Conference Organization
Optimized advertising content delivery in affiliate networks
A holistic Methodology for model-driven B2B Integration: From Business Values over Business Collaborations to Deployment Artifacts
Multi-Agent Decision Support System for Supply Chain Management
A Holistic Approach Towards a UML Profile for Business Modeling
A Multi-Agent System for Content Trading in Electronic Telecom Markets Using
Making Pre-Trip Services Context-Aware
ABSTRACT The process of selection, configuration and consumption of tourist information services is a complex task for the user. This is not least since existing tools most often focus on supporting either the pre- or post-trip phase or the on-trip phase itself by providing context-aware services. The goal of this thesis is to establish a framework that makes pre-trip services context-aware, thus reducing the gap between the pre-trip and on-trip phase by providing a single point of access. This is done by facilitating service selection and configuration in the pre-trip phase and context-aware service consumption in the on-trip phase. Its applicability and feasibility will be proved by a working prototype and evaluated through field studies.
Categories and Subject Descriptors H4.m [Information Systems Applications]: Miscellaneous General Terms Algorithms, Design, Human Factors.
Keywords tourist life cycle, customization, mobile tourist services, contextawareness 1. Introduction Tourism is an information intensive business. Since tourism products are virtual products prior to consumption, travelers depend heavily on tourism information. In the ideal case, tourism services should support tourists with travel-related information during all phases (pre-trip, on-trip, post-trip phase) of the tourist life cycle [9] (cf. Figure 1).
In the pre-trip phase, tourists need information for planning purposes and decision making. After their trip, focus is on reminiscing about the journey and sharing the gained impressions and experiences with friends. In the on-trip phase, however, tourists are mobile and act in unknown environments where they would especially need personalized on-trip assistance in the form of information about accommodation, points of interest (POIs) (e.g., environmental and landscape attractions or gastronomy), weather forecasts, news or safety issues. Mobile services, i.e., services that can be used independently of temporal and spatial constraints and that are accessed through a mobile handset, may address these issues. They have the task to satisfy information requirements of tourists while being on the move by providing them with a broad range of up-to-date, situation-specific information. This information may be in addition adapted to the current situation of the user by exploiting user preferences, user location as well as mobile handset capabilities.
In the last years, research has been very active in each of the
phases of the tourist life cycle. Research in the pre- and post-trip
phase is closely linked to online travel communities [
The goal of online tourism communities is the provision of
up-todate, freely available tourism-related content, thereby enabling
members to collect, view and exchange data items such as blog
entries or pictures or to add own content and reviews. They
provide good support for the pre- and post-trip phases but fail to
support tourists sufficiently during the on-trip phase. A few
communities such as the Tripadvisor1 or the Yahoo Trip Planner2
enable their users to download or print the personal trip plan, but
do not provide support to access this information in a way suitable
for mobile phones. Support for dynamic, on-the-move
information is rare. Customization, i.e., adapting the information
content towards the current context, is missing at all. As they only
provide services which are useful before and after the trip and
which are not interlinked to the on-trip phase, tourists have to use
other sources to satisfy their information requirements while they
are on vacation. Online tourism communities often provide
personal trip planner tools (cf. e.g., [
Research with respect to the on-trip phase has resulted in a wide
range of mobile tourist guides. Since one of the first famous
prototypes [
In the pre-trip phase, tourists should have the possibility to select and configure those tourist services that appear useful to them later in the on-trip phase. In the on-trip phase, the pre-configured services can then provide personal information that is tailored to the current situation and requirements of the tourist and presented on the mobile device of the user.
In the following, this visionary goal is presented by describing possible scenarios from a tourist’s point of view. The vision might include many assumptions. We want to point out that it is not goal of this PhD to work on all these visionary service descriptions, but some of them will be realized within a working prototype. The envisioned system is called virtual tourist agent (VTA) since it should provide tourists the same comfort as if they would call their personal travel agent at home to satisfy their information requirements.
The show case gives some impression how the support of tourists in the pre-trip phase and during the vacation can be realized.
Pre-Trip Phase
In the pre-trip phase, tourists often do not plan all the activities
they intend to undertake in advance, they rather follow an
optimistic approach. For this, the VTA provides a trip-planner
tool to establish a rough schedule that contains the cities/places
they want to visit within a certain timeframe and the route
between those places. Next, the tourists can identify and select all
the tourist information services that they need later in the on-trip
phase. Possible services include a flight information service, a
tourist attraction service or a weather service. In [
The customisation dimension expresses to which extent the information sent to the user is customized to fit the requirements of the respective user by taking into account various context factors.
Concerning the initiation of delivery, services can be classified into pull and push services. Pull services are characterized by a user-triggered search whereas push services deliver information to the user automatically.
For example, the user can select the weather service and configure it to push (service initiation) the up-to-date weather forecast (service delivery) every morning to the mobile device of the user, filtered to the destination of the user (customization).
Since a mobile device suffers from several limitations, e.g., small screen and network connection with low bandwidth, it is important to limit the amount of information that is presented to the tourist so that he or she can obtain the essential information in a non-intrusive way. After all, the mobile device should function as an assistive tool for the current task (e.g., sightseeing) and should not require full attention of the user. To fulfill this purpose, an automation of service delivery and decision making is required. This is done based on the definition and evaluation of rules that may partly be defined by the tourist. These rules are checked by the VTA in order to deliver only relevant information, leading to a more fulfilling user experience. For example, the tourist can define the rule that the VTA should contact him/her in security-related issues, e.g., he or she booked a flight to a country where an earthquake happened (security issue), so that he or she can decide to re-book the flight to another destination. In case of re-booking, the VTA may automatically cancel the accommodation booking on behalf of the user (if he defined another rule concerning this issue). During the on-trip phase, tourists receive support from the VTA based on the service selection and configuration done in the pretrip phase. This means that they obtain information from all the services that have been configured to act in a push-based manner.
Further, the situation of the tourist (comprising context factors
such as location, time, user profile and travel schedule) is
constantly checked by the VTA to detect reactive situations [
Another example might be the case that tourists plan a bicycle tour on a specific day during their vacation. The VTA informs them that the weather forecast might be bad this day and instead suggest them indoor activities. Another case might be a planned mountain tour, e.g., on the vulcano Etna, situated on Sicily. The VTA can check whether it is allowed to go up on the Etna (it is still an active vulcano and dangerous eruptions may occur). In case that it is not, the VTA can suggest a tour on the vulcanos “Vulcano” or “Stromboli”, being part of the Aeolian Islands in the north of Sicily by exploiting knowledge from a respective knowledge base. If a certain activity cannot be carried out at the destination, based on reasons such as bad weather, closing hours or long waiting lines, the VTA can look up the time frame planned for this activity in the trip schedule and suggest an equal activity instead. This shows that the travel schedule is an important source of knowledge next to traditional context factors such as location, time or user profile. The tourists may also pull for information, such as querying the VTA for a nice beach to go for swimming. The VTA can present a personalized list of nice beaches. 3. State of the Art In the following, we report on state of the art in e-tourism by discussing research on online travel communities that focus mainly on the pre- and post trip phase and research in the field of mobile tourist guides, covering the on-trip phase.
Concerning online travel communities, we evaluated in [
In [
Several surveys on mobile tourist guides have already been published that evaluate mobile tourist guides not in terms of provided services but focusing more on the technical side such as architecture, user interaction or context-awareness. In [8], we presented a comprehensive overview and comparison of mobile tourist guides. The comparison is based on an evaluation framework that focuses on context and adaptation criteria. In this way, this survey explores the capability of mobile tourist guides to provide customized services, i.e., services that can react to the context by adapting the information. The main statements are that most systems use their own content databases and do not exploit the potential of incorporating external content, e.g., through standardized interfaces such as web services. Most approaches only consider location and user profile as context factors, while neglecting other ones, such as time or network. Moreover, the potential of combining context properties to derive more valuable logical information is not exploited. 4. Methodology As already mentioned, the goal of this work is to find a satisfactory solution for making pre-trip services context-aware.
To reach this goal, an approach based on the design-science paradigm (cf. e.g., Hevner et al (2004)) is used. The designscience paradigm seeks to create knowledge and understanding of a problem domain and its solution through the building and application of innovative design artifacts. Thereby, artifacts are defined not only as the resulting instantiations (working prototype), but also comprise constructs (vocabulary), models (abstractions & representations) and methods (algorithms & practices) applied in the development as well. To demonstrate the applicability and feasibility of this work, the VTA system will be implemented as a working prototype. For its development, several other design artifacts are needed that will be outlined in detail in section 5. These artifacts further contribute to the knowledge in the field of e-tourism. The goal of design-science research is to address unsolved problems in innovative ways and to address solved problems in more effective and efficient ways. This goal is targeted by this thesis through an innovative approach that facilitates service selection, configuration and consumption through a single point of access. The effectiveness and utility of the VTA prototype will be assessed by a field study with tourists and compared to other systems that target the pre- or on-trip phase in isolation. 5. Research Contribution The main contribution is the design of a conceptual framework and the prototypical implementation of the VTA system. Thereby, it is not the goal to implement each component from scratch, but to implement the whole system by heavily reusing and combining existing tools. Figure 3 illustrates the architecture of the VTA system. In the following, the architecture is explained by describing its components and listing research tasks that have to be addressed to develop the whole system.
In the pre-trip phase the user can select and configure all relevant service types and choose the adequate service providers from which the data is obtained during the on-trip phase. The service instances abstract from the different application programming interfaces (APIs) of the service providers and are linked over a standardized interface to the respective service type. For example, the service type “weather” can access weather service providers using different technologies, ranging from SOAP and restful web services to RSS feeds and finally to wrappers that extract the data from the provider’s website if an API is not available. The different symbols of the service providers act as representatives for the different technologies that can be used to access the data.
In the runtime phase, the VTA system exploits the situation of the tourist and manages the states of the service providers in order to detect changes that require an action on behalf of or sending some information to the user. The main components of the VTA framework are described in the following. For this, Figure 4 shows a high-level view on the components of this framework. Context-aware information systems have to derive meaningful information based on the situation of the user. Ontologies are a promising technology to model the situation of the user since they can represent the knowledge in a semantically rich kind and are therefore a central part of the system. In the pre-trip phase, the VTA needs to include a trip planner tool that assists in trip planning. A survey has to be done to select a promising trip planner tool that can be integrated within the VTA.
The trip planner of the destination portal of New Zealand6 seems to be a good starting point.
Business Services & Service Registry
In our previous work [
Context Services and Context Manager The framework provides access to different context factors such as location time or profile of the user that are provided by the system or by the user or even external context factors such as weather information. As soon as a context change happens, either triggered by the user (e.g., location change), by the system (e.g., time) or by external providers (e.g., weather), the context manager forwards this event to the rule engine.
Rule Engine The rule engine consists of an inference engine and a dependency checker. The inference engine takes the request from the user and the events from the context manager and matches them with the knowledge base that contains the situation of the user and his/her travel plan. In this way, the inference engine uses axiomatic knowledge in the knowledge base to derive new conclusions. For example, when the context manager sends a bad weather event for the next day, the rule engine checks all outdoor activities taking place next day and apply some actions on them based on predefined rules (such as canceling those actions). The activities in the travel plan show complex dependencies amongst each other. For example, if the user plans to stay at a certain destination for another day in order to attend a concert (new activity), this new activity might be dependent on the possibility to stay another night in the booked hotel. If the user already booked a hotel at another destination for this day, this has to be canceled as well. It is the task of the dependency checker to monitor coherent activities.
Workflow Manager The workflow manager encapsulates all different services as workflow components. The workflow manager receives from the rule engine the information about those services that have to be called in order to perform the actions defined in the rule engine and composes the respective services to a workflow (e.g., cancel hotel, send confirmation in form of SMS to the user). 6 http://www.newzealand.com/ 6. Future Work Future work concentrates on detailing the architecture and the necessary components. Based on the design decisions, e.g., heavy-weight vs. light-weight client, existing tools, middleware and frameworks have to be selected that facilitate the design of the conceptual architecture and the implementation of the resulting prototype. Surveys are needed to select the suitable tools and adapt them in a later step for our work. The system will be implemented based on a rapid prototypical approach, which allows testing the prototype in small, iterative steps in order to get fast feedback for further improving the prototype.
Optimized advertising content delivery in affiliate networks
Daniel Mican Business Information Systems Department
Babe s.-Bolyai University str. Theodor Mihali 58-60 400591, Cluj-Napoca, Romania daniel.mican@econ.ubbcluj.ro ABSTRACT This study tackles the problem of advertising content distribution in a liate networks. We model the a liate network as a new business-to-business relationship in which a master site tries to improve its pro ts by properly targeting the advertising information. In our speci c case, we will optimize discount coupons delivery for one of the biggest coupons site in the world. In our study we will employ adapted recommendation strategies based on collaborative ltering methods. The purpose of the present paper is to correlate the display of advertising information on a liate sites with the actual improvements in sales, as a direct method to obtain the user rating which does not exist.
Categories and Subject Descriptors K.4.4 [Computer and Society]: Electronic Commerce| Distributed Commercial Transactions ; H.3.4 [Information Storage and Retrieval]: Systems and Software|User pro
les Keywords recommender systems, collaborative ltering, a liates, content delivery 1. INTRODUCTION This paper describes the research proposal we are working at and which is part of my PhD thesis. It speci cally targets electronic business by tackling a new and emerging businessto-business relationship. In electronic B2B, a provider has various means of attracting new customers, one of them being to supply and promote discount coupons. A proper platform for spreading out advertising information regarding discounts is required, and the information providers should possess intelligent tools to observe and evaluate the e ciency Daniel Mican is 2nd year PhD student in Business Information Systems at Babes.-Bolyai University of Cluj-Napoca, Romania and his advisor is Prof. Dr. Nicolae Tomai, nicolae.tomai@econ.ubbcluj.ro of the discounts information distribution. A liate programs represent possible solution to be adopted. In e-commerce, an a liate is a website which links back to an e-commerce site like Amazon.com. While in classical e-commerce, the provider can easily compute the e ciency and the e ectiveness of a particular a liate, because it can register all products sold through that a liate, in discount information advertising this operation is no longer possible. In discount advertising, the customer usually buys the product directly from the provider e-commerce platform and the provider has no mean to directly identify how the customer acquired the information about the discount. Thus, classical tools for personalized content delivery can not be directly applied for optimizing the advertising delivery information, due to the fact that the direct feedback loop is missing.
In this paper we will give a brief description of the research problem stressed in the paragraph above, stating out the challenges we are facing. The paper is organized as follows.
Section 2 presents the research problem of my PhD thesis. In section 3 we present the state of the art, mainly concerning the intelligent techniques for content delivery and studies concerning the a liates problem. Section 4 describes the main challenges we face in our study, while section 5 presents the wow-coupons.com use case. We conclude the paper by presenting our future plans. 2. RESEARCH PROBLEM Our main research problem is to build an e-business system in the area of a liates marketing and content delivery.
Speci cally, we try to develop a delivery system for a major coupons site. Coupons represent a well-established way of providing bene ts to customers for a speci c activity to improve their business.
Online coupons or online coupon codes are discounts and bargain deals for all major on line stores and shopping sites on the Internet. Online stores creates discount codes to select groups of customers. By using coupon codes some groups of people will enjoy discounts when shopping online, like free shipping. Discount coupons is a way to save with on line shopping and some codes will be automatically applied at checkout. Other codes must be copy pasted intro a special coupon text box before to con rm the order.
An independent research agency made a recent survey on behalf of Pay By Touch on over 100 shoppers that were questioned in May 2007. The survey proves [12] that shoppers develop a positive response to discount schemes if these are both highly targeted and convenient to use. In fact, the survey reveals that 88% of shoppers would use discount coupons, if these were more focused on their product preferences and were available in store while they were shopping.
Of those surveyed, 95% of shoppers who used a retail loyalty card have received in-store discount coupons. However, 75% of these shoppers said they frequently forget to redeem them even if some of the discounts o ered were on items they normally buy. Other factors that contributed to low redemption rates were the inconvenience of having to carry pieces of paper around and the fact that discounts were mostly for items the shopper did not have a history of buying. Therefore, it seems clear that retailers who o er targeted discounts and make it easy for those o ers to be redeemed, could have a compelling way to develop customer loyalty and increase pro t.
In our study case we want to speci cally develop and implement a content delivery system for one of the top coupons site in USA (http://wow-coupons.com/). From now on, we will name this site as the master site. Our system will search, select and deliver coupons on behalf of 3'rd party sites that sell or recommend products, or for community sites that offer valuable user targeted content. Delivery will be targeted to a network of a liates, established in order to enhance information dissemination. Delivered content will assist the site visitors, the community members in their online shopping activities and will help them to save money when they buy products. Our main target is to optimize the system such us to deliver highly targeted discount coupons. For our target, we intend to adapt recommendation systems and collaborative ltering technologies. Another feature of our interest is to technically enable the coupons site to deliver the content in real-time, i.e. without the usage of o -line information exchange with the a liates. When a new coupon is added on the master site, the content delivered on the a liate sites should be seamless updated.
Collaborative ltering is of our interest because it is a way to establish what items to display to web users who have browsed some ads or made purchases in the past. Collaborative ltering software compiles purchasing information on customers to pool them into clusters and uses some cluster members' purchasing patterns to predict the buying habits of others in the same cluster. It does this in real time and, for instance, puts an ad on the customer's screen while he or she is making a purchase [10]. In our case we want to display targeted discount coupons to users that are going to make a purchase. For example, we target an online shopping website that makes discount recommendations for the products the customer has in her shopping chart. This will help her to save money and also to develop a strong loyalty for that online shop because it proves that the shop cares about the customers and their needs.
To directly apply collaborative ltering, we need to collect and represent di erent partner sites and people preferences.
Preferences are typically based on item ratings (i.e. posteriori feedback) explicitly delivered by users. The system recommends products which were evaluated positively by another similar user or by a set of such users.
Our main challenge is that we cannot collect the user ratings or preferences because we deliver the advertising content (coupon codes). The coupon code for a particular product and promotion campaign is the same, regardless the site that publishes it. When a user applies a coupon by buying a product, we do not have a direct method to identify the a liate site that published the coupon to the user. Therefore, in our business setup, we do not have the feedback loop mechanism to allow us to use a classical collaborative
ltering method.
Instead, we want to implement a module part of the master site system that will allow it build a pro le for each a liate. This module will track down and monitor how many times an a liate user viewed and clicked on every coupon displayed. On the other side, the master knows the number of products sold per every coupon delivered to all a liates.
All these will aggregate in a statistical information that will replace the classical feedback from the partners and will provide the a liate pro le. The master can optimize and improve the coupons delivery algorithm for the entire a liates network.
In this section we described our research problem in the terms of our B2B relationships. But, the problem is more general, spanning over all business relationships where direct feedback or user rating is not available. By tackling the a liates problem, we intend to prove that intelligent techniques for content information delivery are more suited and can enhance the pro t of the advertisers, even if there are only statistical glues about the real pro le of the information users. 3. STATE OF THE ART IN THE FIELD In this section we will investigate the state of the art in the elds covered by our research. We should emphasize that there is a lot of bibliography tackling recommender systems and collaborative ltering. Regarding the a liates, there are a lot of business studies concerning the a liates in e-commerce setups. In e-commerce setups, the feedback loop is closed because the master can identify through which channel a speci c purchase was performed. Therefore, besides a proper advertising, one of their biggest challenge is how to compensate or reward each a liate, accordingly to the pro ts they generated. In our study we are not interested in designing a proper reward scheme for a liates, we suppose that we have good mechanisms to maintain the network and make it growing. 3.1 Recommender systems and collaborative
filtering
Recommender systems are an important part of recent
ecommerce. They enable the increase of sales by suggesting
to users selected products on o er. The problem of how to
choose the most suitable items, possibly with respect to the
user's inclinations, is a challenging research problem that
has been investigated for many years [
Four fundamental approaches to recommendation can be
mentioned: demographic ltering, collaborative and
contentbased recommendation, and simpli ed statistical approaches
[
In demographic recommendation, users are classi ed based on their personal data, which was collected during the registration process, survey responses or other feedback methods.
Each product is assigned to one or more classes with certain weights and the user is attracted to items from the class closest to their pro le. This is attribute based recommendation.
Collaborative recommendation is typically based on item ratings explicitly delivered by users. The system recommends products, which have been evaluated positively by another similar user or by a set of such users, whose ratings have the strongest correlation with the current user. This is user-touser correlation.
Content-based recommendation focuses on the similarity between products, usually taking into account their features like textual descriptions, hyperlinks, related ratings, or cooccurrence in the same purchased transactions or web user sessions. Items that are the closest to the most recently processed (viewed), are recommended regardless of user preferences. This is item-to-item correlation. Association rules and sequential patterns are the most interesting techniques used in recommendation based on item-to-item correlation.
They are usually applied to data sets related to items such as purchases, ratings of TV programs, navigation paths rather than directly to item attributes.
In the statistical approach, the user is shown products based on some statistical factors; usually popularity measures like averages or summary ratings (the best rated), and numbers of sold units (the best buy) [9].
The information overload problem a ects our everyday experience while searching for knowledge on a topic. To overcome this problem, we often rely on suggestions from others who have more experience on the topic. However, in web case where there are numerous suggestions, it is not easy to detect the trustworthy ones. Shifting from individual to collective suggestions, the process of recommendation becomes controllable. This is attained with the introduction of Collaborative Filtering (CF), which provides recommendations based on the suggestions of users who have similar preferences. Since CF is able to capture the particular preferences of a user, it has become one of the most popular methods in recommender systems [11].
A web site or other online service that receives extensive tra c has the potential to analyze the resulting usage data for the bene t of its user population. One of the most common applications of such analysis is collaborative ltering.
A web site o ering items for sale or download can analyze
the aggregate decisions of the whole population, and then
make recommendations to individual users of further items
that they are likely to be interested in. The
recommendations made to a speci c user are thus based not just on his
or her own previous actions, but also on collaborative
information, the information collected from other users in the
system [
Collaborative ltering algorithms can be categorized as [
User-Based algorithms: operate on the assumption that consumers who have bought similar products in the past will prefer to buy similar products in the future Item-Based algorithms: operate on the assumption that items that have been co-purchased in the past will continue to be co-purchased in the future Regarding the technical mean of delivering the recommendations, CF are split in [8]: memory-based algorithms, which recommend according to the preferences of nearest neighbors. They utilize the entire user-item database to generate a prediction. These systems employ statistical techniques to nd a set of users, known as neighbors, that have a history of agreeing with the target user (i.e., they either rate di erent items similarly or they tend to buy similar set of items). Once a neighborhood of users is formed, these systems use di erent algorithms to combine the preferences of neighbors to produce a prediction or top-N recommendation for the active user. The techniques, also known as nearest-neighbor or user-based collaborative ltering are more popular and widely used in practice. model-based algorithms, which recommend by rst developing a model of user ratings. Algorithms in this category take a probabilistic approach and envision the collaborative ltering process as computing the expected value of a user prediction, given his/her ratings on other items. The model building process is performed by di erent machine learning algorithms such as Bayesian network, clustering, and rule-based approaches. The Bayesian network model formulates a probabilistic model for collaborative ltering problem. Clustering model treats collaborative ltering as a classi cation problem and works by clustering similar users in same class and estimating the probability that a particular user is in a particular class C, and from there computes the conditional probability of ratings. The rule-based approach applies association rule discovery algorithms to nd association between co-purchased items and then generates item recommendation based on the strength of the association between items.
Both practical experience and related research have reported that memory-based algorithms present excellent performance, in terms of accuracy, for multivalue rating data. On the other hand, model-based algorithms are e ciently handle scalability to large data sets [11]. 3.2 Affiliates marketing A liate marketing is a web-based marketing practice in which a business rewards one or more a liates for each visitor or customer brought about by the a liate's marketing e orts. A merchant, also known as an advertiser or retailer, is a web site or company that sells a product or service online, accepts payments and ful lls orders. A liates (also called publishers) place merchants' ads, text links, or product links on their web sites, shopping engines, blogs, etc. or include them in e-mail campaigns and search listings in exchange for commissions on leads or sales. 4. MAIN CONTRIBUTIONS EXPECTED We expect to contribute in several main areas.
From the systemic point of view, we intend to de ne the overall picture comprising all business to business relationships. Figure 1 presents an overall sketch of our system. Our contribution will mainly reside in the WOW-Coupons.com master site where we will technically embed a recommender system.
From a systemic point of view, we intend to de ne the overall picture comprising all business to business relationships.
Our contribution will mainly reside in the WOW-Coupons.com master site where we will technically embed a recommender system.
From a technical point of view, a challenge is how to pack
the content delivered to a liate sites so as the partners will
seamlessly accept the content and publish it into their
websites. The content should be delivered in such a way that
we will be able to register the number of views per coupon
and number of clicks per coupon. We envisage the usage of
server-controlled web development techniques that exchange
small amount of information behind the scene. AJAX
technologies [
From a algorithmic point of view, our research raises several open issues: to develop a statistical analysis tool in order to relate in a signi cant manner the collected information about content display (number of views and clicks) and the received payments from the earning reports. Statistical analysis will provide a way to collect the required feedback in the collaborative ltering methods to select a proper user pro le management scheme.
This is highly related with adapting a recommendation system or collaborative ltering method to our setup.
Up to now, we only investigated several alternatives, but we did not established which algorithm would be more suited for us. We expect the ICEC doctoral symposium to allow us to progress on this issue, by having discussions with highly esteemed experts to compare the intelligent content distribution to the actual functioning of the a liates system. Actually, there is not too much intelligence in the system, in the sense that the master delivers to each site the coupons related with the latest marketing campaign added in the master's database. In order to allow for a nonbiased comparison, we need to design a controlled experiment on wow-coupons.com, covering a de ned period of time and a stable part of the a liates network.
Designing and running the experiments are a big challenge, because, in meantime, we need not to downward the business performance over the period of the experiment. 5. WOW-COUPONS.COM We started this research because we intend to improve the way that WOW-Coupons.com delivers the coupons for the a liate sites. The actual system delivers the last coupons added on the master site, organized by store and by category.
Up to now, this delivery scheme proved to be winning, but in nowadays this module needs some improvements because we intend to enhance the user usage of time and also the pro ts wow-coupons.com makes from the a liates network.
We will try to do a brief description of the site that will use the recommender system that we want to provide below.
The overwhelming majority of consumers in the US collect coupons to help them make the most of their money. Consumers spend plenty during the holiday season, and with tightening budgets, they are spending money where they see the best deals. During the holidays, an impressive 71% of survey respondents [12] say they're likely to use the Internet to research and compare holiday products and gifts.
WOW-Coupons.com is the fastest-growing coupons site on the web. It is a winner of 2005 LinkShare Gold Link "Innovative A liate" award. Averaging around 600,000 unique visitors a month and with a steadily growing community of e-mail newsletter subscribers, currently more than 80,000, it commands positions at the top of the search engines.
A few time ago the company launched the UK version of the site WOW-Coupons.co.uk. Categories include printable vouchers with Printable Retail, Printable Grocery, Printable Restaurant and Printable Travel Coupons subcategories, as well as Online Coupons.
Printable retail coupons section will provide real coupons buyers can take to favorite major retailers and national franchise stores. Shoppers choose the discount, print it out, and go shopping without needing to sign up. The process eliminates clogged inboxes with endless promotional e-mails from dozens of mailing lists. Grocery printable vouchers section will provide the grocery coupons users need, when they need them. Print coupons to save on favorite brands at supermarkets and drugstores everywhere. Restaurant printable vouchers section will provide coupons that can be printed out and used to save money at restaurants across the nation. When eating out may seem like an extravagance, think again and enjoy dinner. Travel and entertainment printable vouchers section will provide discounts and vouchers for diverse traveling destinations whether in US, U.K. or internationally. Travelers can still a ord to take vacations or visit the family. The site has printable travel coupons and o ers for car rentals, accommodations, amusement parks, museums and lots more.
Just like printable vouchers, the best and biggest retailers and service providers o er online coupons and special discounts. The di erence is, these can be used only for purchases made online. On the site visitors can browse great o ers, go straight to a favorite store (arranged alphabetically or by date posted) to see the latest deals, or to use the navigation menu on the right side of the site. When an item is found, the user can follow directions in the description of each coupon to be sure that savings have been applied before paying for an order.
For implementing the architectural design for coupons
delivery, we employed the XML standard to provide an RSS
[
les that can be used on the Internet. The system generates an XML le and update it hourly for every company that have coupons in the database. 6. FUTURE PLANS Our future plans intend to cover the above mentioned issues.
We are only in the incipient stage of our research, because I have spent all the rst year in taking exams and now, we worked on de ning a proper problem and researching various technologies for this problem. Up to now, Wowcoupons got implemented without the optimization feature and to perform the optimization mentioned in this paper is a good plan for future research. We also think that this study will represent a good opportunity to advertise to the a liate marketing community the potential of intelligent techniques. 7. ACKNOWLEDGMENTS This work is supported by the Romanian Authority for Scienti c Research under project IDEI 573. We also acknowledge the support from Mrs. Elena Potoupa, CEO of WOW Things Inc., owner of wow-coupons.com.
A holistic Methodology for model-driven B2B Integration: From Business Values over Business Collaborations to Deployment Artifacts
Marco Zapletal
Electronic Commerce Group Institute of Software Technology and Interactive Systems
Vienna University of Technology
marco@ec.tuwien.ac.at ABSTRACT Business-to-business (B2B) electronic commerce based on the principles of Electronic Data Interchange (EDI) systems has been conducted for a long time. In recent years, we observe a switch to a rather business process-based thinking for implementing interorganizational systems. UN/CEFACT’s Modeling Methodology (UMM) - which I co-authored - is considered as one of the mature graphical modeling approaches for modeling interorganizational business processes. However, UMM has still several shortcoming that prevent a throughout model-driven software engineering approach. In this PhD proposal, further contributions to the UMM are identified that are required to lift it to an holistic B2B methodology for the development of interorganizational systems. These contributions will extend the UMM to become an integrated approach starting with business models, leading over to business collaboration models, and finally resulting in deployable artifacts for business service interfaces. The proposed top-down approach is in line with the ideas of model-driven engineering resulting in shorter development cycles and reduced complexity. 1. MOTIVATION Conducting electronic business between enterprises was not an invention of the internet age, but has existed for decades. However, requirements of business-to-business (B2B) electronic commerce have changed since that time. In former days, when B2B electronic commerce was referred to as Electronic Data Interchange (EDI), its focus was document-centric. This means, in order to avoid bilateral agreements on business documents, business partners agreed on business document standards. But, as history has shown, the results of these standardization efforts were overloaded and ambiguous document standards. This led to costly EDI systems and participation in electronic business was reserved to large companies that were able to afford such implementations. As a consequence, only circles of acquainted enterprises exchanged business 2nd year PhD student at Vienna University of Technology messages electronically in order to reach their business goals and gain financial benefits.
With the advent of the internet, the area of electronic business started to boom. In the field of B2B electronic commerce, small and medium sized companies now saw their chance to enter electronic markets.
Now, it seemed possible to find new business partners electronically and to dynamically conduct e-business. In addition, with the advent of XML, the problems of EDI appeared to be solved all of a sudden. However, this was a broad misconception - the pure mapping of EDI concepts to brackets did not yield a solution to the shortcomings of traditional EDI.
At this time, business process management was already in use to specify intraorganizational workflows. Enterprises started to adopt business process modeling in order to monitor their procedures and to design process-based solutions. In the context of EDI, the concept of a business process has already existed - but buried in the minds of those people that were responsible for the interorganizational systems. These people were aware, for example, what to do next when an invoice was received and how to trigger manual compensation if - in case of a failure - a dunning letter was received before an invoice. They were able to resolve the problem by phoning the business partner, because their counterpart was known to them. In this respect, the notion of a business process - as a protocol for specifying the course of business - was already there.
However, according to the idea of modern electronic markets where companies of almost any size conduct business in a dynamic way, business partners are not acquainted as described above. Dynamic B2B e-business involves spontaneous agreements, which might exist just for one economic transaction. There are no offline negotiations and no face-to-face relationships. Instead, agreements are made online, which requires business partners to unambiguously define how to conduct business with them. In other words, business partners must describe what business processes they offer in order to show potential business partners how to interact with them. It follows, that interorganizational business process models are the basic building block for flexible and spontaneous B2B ecommerce.
Figure 1 shows a slightly extended version of the Open-edi reference model [14]. It separates the development of interorganizational systems into business and technology concerns. Specifications for capturing collaborative business logic are covered by the business operational view (BOV). The functional service view (FSV) comBusiness Process Models
Deployment Artifacts
BOV
FSV prises technology specifications for implementing business logic.
In this thesis a holistic approach for B2B integration is developed starting with business models, leading to business process models, and finally resulting in deployable artifacts.
Modeling interorganizational business processes is considered as
the foundation of this approach. Hence, the currently most
promising approach for modeling interorganizational systems
UN/CEFACT’s Modeling Methodology (UMM) - will be the
starting point of this thesis. I co-authored UMM 1.0 during my
undergraduate studies [15] and was one of the authors of the first book
on UMM [
The remainder of this proposal is structured as follows: Section 2 describes the state of the art in regard to current B2B modeling approaches. Section 3 elaborates the contribution of this work by outlining solutions to current shortcoming of the UMM. Each shortcoming as well as the corresponding solution is discussed in its own sub section. Finally, section 4 concludes this proposal. 2. STATE OF THE ART People have learned that traditional EDI concepts do not realize the idea of dynamic e-commerce as envisioned at the beginning of this proposal. In this respect, the need for modeling interorganizational business processes has become evident.
Traditionally, business process modeling focused on intraorganizational business processes in order to capture workflows that are internal to an enterprise. Internal processes are always modeled from the perspective of the respective company. In a collaborative context, however, a partner-specific view on a process is not sufficient. If each participant in a collaborative process describes its own perspective on the same process in isolation, the resulting process descriptions will most likely not match. Thus, modeling interorganizational processes requires a global perspective.
Today, several modeling approaches exist for capturing
collaborative business processes. Some appropriate approaches have been
identified in [
UMM defines a meta model and a development process ranging
from requirements elicitation to business collaboration design. In
previous versions of the UMM - before version 1.0 - there was a
lack of formal correctness of the meta model, which made it
impossible to derive software artifacts from the model according to
a model-driven software development approach. In addition, the
meta model’s complexity was the reason that applying the UMM
was a tremendous task - oftentimes it resulted in faulty business
collaboration models. The current version 1.0 of the UMM [
The current UMM 1.0 will be the starting point for this thesis. On top of UMM, an integrated B2B methodology following three layer approach will be defined as outlined in the section before. A detailed description of current problem fields and the contributions of this thesis are given in the next section. 3. PROBLEM FIELDS - CONTRIBUTION OF
THIS THESIS
When I started to work on the UMM, it was a so-called "UML
profile" but it was not formally specified as one. A UML profile
customizes UML for a domain-specific purpose by defining a set
of stereotypes, tagged values and constraints. The versions before
UMM 1.0 [
ology to UML 2 The UMM Foundation Module 1.0 is the first UMM version that satisfies the formal requirements of a UML profile. When the UMM 1.0 project was started within UN/CEFACT, UML 2.0 was not considered as stable enough. Hence, UMM is currently built on UML 1.4, but today UML 2 is considered as the state of the art. Consequently, UMM stakeholders ask for a "UMM 2.0" that is defined on top of UML 2. In addition, UML 2 provides major improvements to key modeling elements of the UMM (e.g., activity diagrams). It follows that moving UMM to UML 2 is required. Beside the criticism that the current UMM is based on an outdated UML, the meta model of the UMM is still often bashed as too complex (e.g., UMM 1.0 models often result in excessive package structures). Furthermore, some workarounds that were necessary in the meta model due to the use of UML 1.4 contribute to often bloated UMM models.
For this reason, the initial contribution of this thesis will be a
definition of UMM on top of UML 2, reflecting experiences and
comments from stakeholders. The result is an easier to use UMM that
builds on current standards. This ensures further adoption by
potential users and fosters the support of tool vendors. The new UMM
version serves as the core for the further extensions and
improvements suggested throughout this thesis. The use of UML 2 also
eliminates the above mentioned workarounds in the UMM meta
model. In [
Beside UML activity diagrams several other notations and
modeling languages emerged in the past to capture process flows. In the
past, the Business Process Modeling Notation (BPMN) [
gineering into UMM
Modeling business processes shows how an enterprise acts in
order to reach an economic goal. In the context of B2B, the
business process model captures how different enterprises interact to
exchange objects of economic value. However, a business
process model does not concentrate on economic reciprocity - what
objects of value are exchanged to gain other objects of value. It is
the purpose of business models to capture this aspect of economic
transactions. According to Timmer [
The current UMM concentrates on specifying business process models as well as their requirements, but lacks value-based requirements engineering by means of business models. However, it is an interesting fact for an enterprise to combine these different views on economic transaction. It helps identifying business processes that have to be supported in order to realize a given value exchange.
In addition, it allows monitoring if a deployed business process still fulfills a given business model.
In order to provide value-based requirements engineering in UMM,
this thesis will propose the integration of an e-business modeling
approach. As identified in [
The e3-Value approach currently defines its own notation. A first
step towards the integration of e3-Value into the UMM is
definition of a UML profile for e3-Value. This contribution has been
published in [13]. Still an issue to address is aligning the e3-Value
concepts with the UMM development process. In addition,
specifying that a business model fits to a certain business process
models (or vice versa) necessitates consistency checks between those
artifacts. In [
models According to the idea of model driven software development the derivation of deployable artifacts from collaborative business process models is desirable. In the field of Web Services, the Business Process Execution Language (BPEL) [18] gained a lot of attention for implementing business processes. BPEL describes a business process from a partner-specific view. In contrary, UMM focuses on global choreographies. Thus, in order to generate BPEL artifacts from UMM, global UMM business collaboration models have to be mapped to partner-specific BPEL processes. Starting with a graphical model showing a global perspective provides major benefits for three reasons: Firstly, the business collaboration model serves as a kind of contract partners agree on. Secondly, the business collaboration model allows the generation of complementary process specifications for each partner’s interface. This ensures that the partner interfaces interact according to the global choreography of the business collaboration. Finally, the generation of such artifacts allows quick and cheap customizations of a B2B system to changing business requirements.
There already exists some work in the field of deriving deployment
artifacts from business process models. In [
This approach was implemented in [
Besides the pure Web Services approach, this thesis will also focus on the implementation of UMM processes using workflow framworks. The Windows Workflow Foundation (WF) is an upcoming technology allowing developers to create workflow-centric applications. The WF approach is not limited to workflows internal to a company, but allows the implementation of interorganizational business processes. Similary to the BPEL approach, this thesis will propose a derivation algorithm for generating Windows Workflow artifacts from UMM business collaboration models. 3.4 Managing UMM artifacts within business
registries The vision of dynamic B2B presupposes that business partners find each other electronically based on the descriptions of business processes and of the services they need and offer. This idea requires that business partners are provided with means to publish as well as to consume information related to their business conditions. It follows, that the concept of a business registry is required in order to provide enterprises with a central site for to find each other.
Such a business registry must be capable of managing UMM artifacts. Artifacts might be whole models or just parts thereof. Parts of a UMM model might be re-used in another interorganizational business process.
In order to solve the issue explained above, this thesis will describe the representation and management of UMM artifacts within business registries. More specifically, ebXML registries will be the target platform for managing UMM artifacts. The work in terms of registering artifacts is made up of two parts: Firstly, approaches are proposed to store UMM models as a whole or just several parts thereof. Since dependencies might exist between different parts of a UMM model, the proposed approach will outline how to maintain these relationships in a business registry.
We already published this contribution in [10] and [9].
Secondly, since business models and UMM business collaboration models are combined a representation of business models within a business registry is required. The thesis will include an approach describing semantic links between business models and business process models. This enables potential business partners to find each other based on business models or business process models.
This contribution was published in [12]. 4. CONCLUSION The goal of this thesis is a holistic B2B modeling methodology potentiating enterprises to participate in dynamic e-business environments. The current UMM will be revised and amply extended in order to be a tool for enabling successful and real-world B2B integration. The resulting methodology follows a three step top-down approach starting with business models, leading to business collaboration models, and finally resulting in deployable artifacts for business service interfaces. Applying the methodology will foster the vision of dynamic B2B e-commerce.
Multi-Agent Decision Support System for
Supply Chain Management
Yevgeniya Kovalchuk Department of Computing and Electronic Systems
University of Essex +44(0)1206 87 3805 ABSTRACT This paper presents an extended abstract of the author’s doctoral research project on developing a multi-agent intelligent system for automatic managing supply chains. Supply chain management (SCM) is a very complex and dynamic environment. The doctoral work, which started in October 2005, is dedicated to finding better solutions for successful performance in the domain of real-time SCM.
Categories and Subject Descriptors I.2.6 [Artificial Intelligence]: Learning H.4.2 [Information Systems Applications]: Types of Systems – decision support General Terms Economics, Algorithms, Design, Experimentation 1. INTRODUCTION While running their business, enterprises usually deal with a number of activities, such as: procurement, production, warehouse management, selling, marketing, and customer servicing among others. To help them to manage these activities, organisations try to automate their business processes. Usually, independent software and hardware solutions are used for each of the activities. However in practice, all the activities are highly connected and interdependent. To integrate some of them in a single process is the task of supply chain management (SCM).
The SCM is concerned with negotiating with suppliers for raw materials, competing for customer orders, managing inventory, scheduling production, and delivering goods to customers. In addition to its complexity, the SCM is also a time-constrained and ever-changing process, especially nowadays, when enterprises move their business on-line. Taking into consideration market globalisation, companies often run distributed businesses, having suppliers and customers all over the world. To deal with their contractors, organisations use the Internet to participate in electronic commerce, where business occurs very fast. To be able to react to all changes quickly, companies are looking for applications that can support dynamic strategies and adapt to new conditions in the environment. The development of such an intelligent decision support system for SCM is the main objective of the author’s PhD project.
Although the aim is to develop an integrated application for SCM, due to its complexity, it is difficult to address all the issues which can arise in the domain of SCM. To narrow the research scope, the project is mainly focused on the demand part of the supply chain. In particular, different methods for predicting customer offer prices that could result in customer orders (winning bidding prices) are explored and compared in the system. The motivation is that expected findings not only can improve a company’s performance while running its supply chains, but could also be applied to financial markets and online auctions where the task of predicting winnings bidding prices is crucial. The TAC SCM game, where software agents developed by different research groups can compete against each other in the context of the SCM, is used as a test bed to evaluate the proposed algorithms. This simulated environment was implemented by Carnegie Mellon University and the Swedish Institute of Computer Science (SICS) in 2003 as part of the International Trading Agent Competition (http://www.sics.se/tac/). The game is now probably the best vehicle for testing SCM systems as it encapsulates many of the tradeoffs that could be found in real SCM environments: timeconstraints, network latency, unpredictable opponents, etc.
The rest of this paper is organized as follows. The description of
the TAC SCM scenario and overview of related work are
provided first. Then, the research approach followed is
presented. The results achieved so far along with the plans for
future work are given next. The paper closes with the
conclusions.
2. THE TAC SCM SCENARIO
According to the TAC SCM scenario [
Semantic based Project Management
A-1040 Wien +43 (0)1 58801-18881 ABSTRACT In the past and still today, projects miss their goals or are cancelled because of overruns in time and budget. Reasons for their failure are that information often gets lost or that it is hard to remember how and where to find the needed information.
To improve this situation it is intended to enhance a project management system with semantic technologies, such as ontologies and semantic search. For access via client this project management system will enhance a fat-client on the Semantic Desktop system. The usage of this semantically enhanced project management system will be demonstrated by a prototype.
Categories and Subject Descriptors H.3.2 [Information Storage and Retrieval]: Information Storage General Terms Management Keywords Project Management, Semantic Desktop
Semantic technologies,
ontologies, 1. INTRODUCTION In the past and still today, projects miss their goals or are cancelled because of overruns in time and budget. Reasons for their failure are that information often gets lost or that it is hard to remember how and where to find the needed information.
The idea to enhance project management systems with semantic technologies enables for example a better search and reuse of already existing data. Thus it reduces time and costs and in addition, it reduces the effort of project management and increases the probability of project success.
To enable a better comprehension of this topic chapter two specifies the research problem and the benefit and chapter three the economic relevance. Chapter four describes the state-of-theart of current Project Management and Semantic Desktop systems and the innovation of this topic. The proposed technical solution will be explained in chapter five, while chapter six gives an overview of the future work. Last but not least chapter seven gives an overview of the scientific contributions and chapter eight gives a short conclusion of this PhD thesis. 2. Research problems The main goal of this PhD work is to enhance Project Management (PM) with semantic technologies and further, to enable an interchange of information between a Semantic Desktop and a Project Management System.
By combining semantic technologies and PM systems, a reduction of the administration effort and an improved control of the progress of a project seem possible. Introducing semantic technologies, such as ontologies, semantic annotation of content and semantic search open up new ways of delivering the needed insight and experience of past projects. Relevant information of former projects is consolidated in a knowledge base. With the use of ontologies project members can search for concepts and do not have to search for exact keywords. Furthermore, different information items are set in relationship which simplifies and optimises the search process. All information items for a project are on one platform or at least a relationship exists between the project data on a desktop and on the platform related to a project.
With the availability of the knowledge of already finished, running and planned projects the probability to deliver a project in time, in budget and with the specified capabilities is improved.
For a better understanding what the expected added value of this PhD topic is, a small Use Case is given in the following: There is a large company, which has lots of different (past, current and future) projects. Very often, new projects are similar to already finished ones or problems might appear in current projects, which were already solved in similar projects in the past.
In those cases, it is necessary to look up again the information, such as documents, contact information or statements of costs of those finished projects.
A project member of a big company is involved in a project where some problems have appeared. This person remembers that similar problems arose in another project a few years ago. To save money and time he wants to get the information of this project and therefore searches for information about it. But he doesn’t know the exact terms to get the description of how the problem was solved in the last project, the name of the person who solved it or wrote the documentation. Current project management tools only allow to search for keywords, different information items, and full text search or maybe in time categories. Hence, the search will be quiet difficult and of course will need time.
The goal of this PhD is to tackle this problem. The user can search for the name of the producer of the document, the creation date, the milestone in which the document is part of, an email where the document is attached or for a name of a project member. The result of this search is presented as an ontology tree that makes it possible to navigate through the possible results until the needed information is found. A normal listing of documents or information items is included in addition, but the main issue is to navigate through different information items based on the semantically annotated data until the user can find the needed information. In this way, the PhD work contributes to research in project management systems. Currently, the user has to know what he/she is looking for. Semantic annotations allow connections between items and the establishment of an ontology so that users can navigate easily through all information items.
In addition, an interface between a Semantic Desktop system and the PM system shall be built, thus enabling an up-to-date access to the relevant information of a project. With this interface, relationships between data and information items of projects on multiple systems and information items on a desktop are established. Hence, also the desktop can be searched for project relevant information.
Another benefit of this thesis is its domain independencies.
Project Management is part of every project, independent if it is part of a health care or of an e-commerce project.
To enhance an existing PM system with semantic technologies the following sub-goals are relevant:
Definition of a Project Management Ontology and PM related Ontologies
Domain For the PhD topic a project management domain ontology as well as other related ontologies have to be developed. These ontologies concern project related issues (e.g., milestones, tasks), project documents, temporal issues, or project members and their capabilities. All these ontologies are not on the same level, but they support and complement each other.
Interface Semantic Desktop – Semantic PM An interface between the client, a Semantic Desktop system and the semantic PM system must be implemented. The Semantic Desktop system must be extended with a fat-client with the functionalities to search and to set links to PM information as well as to edit them. The ontology of the Semantic Desktop must be adapted with parts of the project management ontologies.
Flexible Architecture of the System To guarantee a positive result of this PhD a detailed architecture of the system has to be designed. This architecture has to be built in a modular and flexible way to allow future extensions. Also all interfaces of this system must adhere to open standards.
Proof of Concept of the PM System The proof of concept of the PM system includes a prototype of the system with the following parts: ontologies, databases, semantic technologies (metadata, tagging …), interface to the Semantic Desktop, adaptation of the Semantic Desktop and the functionalities of PM. It also includes the evaluation of the prototype. The PhD deals with project management in general. But to receive useful test results, the prototype directs its attention to project management for IT projects and in greater depth to the tourism (industry) domain.
The main problems of the PhD will be
The evaluation of existing Project Management
Systems One issue of the PhD work is to find an OpenSource Project Management System, based on a client-server application.
Further consideration regards the types of documents that are to be integrated in the system as well as their storage.
Another question is if the system will support automatic annotation of metadata or if the users have to annotate their inputs by themselves. To address these risks, an in-depth evaluation of existing OpenSource PM Systems as well as of possible semantic systems, such as document management or knowledge management systems, will be done. •
To develop ontologies and combining them A major problem is to develop the different types of ontologies, such as a domain ontology for project management and a time and date ontology for time data, and then to link them. The risk here is that the matching does not fit. To avoid these problems an evaluation of possible existing ontologies will be done. Based on these experiences, possible problems should be avoided. A further preliminary consideration regards whether the ontologies should be merged. The PhD work will start with the assumption that they will not be merged, but properly interlinked. •
To ensure valid test results A problem at the end of the work could be poor test results in case of poor test scenarios and questionnaires. To avoid this problem, requirements and use cases will be done at the beginning of the project. In addition, useful test scenarios are needed. Therefore the main tests will be at the Electronic Commerce Group of the Institute of Software Technology and Interactive Systems at the Vienna University of Technology. During and after using the prototype (for a few weeks) questionnaires and test scenarios concerning the usage will be carried through and analyzed. 3. Economic relevance Today many projects miss their goals or are cancelled because of overruns in time and budget. The Chaos Report [10] identifies that 31% of IT projects were cancelled or never completed and nearly 53% of the projects cost almost twice the initial estimate.
Reasons for their failure are that information often gets lost. It is hard to remember how and where to find the needed information.
The working prototype enables a better search and reuse of already existing data by using semantic technologies, reducing thus time and costs. In addition, it reduces the effort of project management and increases the probability of project success.
Information of already finished projects can be used as input for planning of new projects and to monitor progress and risks of projects underway. That is because of better storage of project relevant information. Due to explicitly describing the relationships between various project documents in machine accessible form, better administration of projects and easier identification of relevant information is possible. Based on improved relationships between different information tasks, connections between previous and current projects are possible and a better administration of projects is feasible as well. Another advantage is a faster reaction to project changes because experiences and problems of previous projects and their management are traceable and visible.
Since project management software is a high volume market, and since the solution envisaged tackles the major problems in project management, the economic potential is very high. 4. State-of-the-art This chapter is divided into three parts, the state-of-the-art of Project Management, of ontologies and of Semantic Desktop systems. All of them include the innovation aspects of this PhD thesis. 4.1 Project Management There are several Project Management software tools, which help companies to manage IT projects. Currently, there are no solutions in the field of project management which contain semantic technologies. The available systems differ considerably with regard to their functionality. The spectrum ranges from just time planning (milestones) or resource planning to tools that assist the entire project management process, as for example projectopen1 or dotproject2.
Problems with projects and their management are often a mix of information storage (PM platforms, personal computer …), neglecting of milestones, costs, etc. and inconsistence of stored information (versioning of documents). These are only some reasons why projects might suffer overruns in time or budget or miss their goals or are cancelled. The Chaos Report of the Standish Group [10] identifies that in 1994 US companies invested approx. $250 billion in IT development. 31% of the projects were cancelled or never completed and nearly 53% of the projects cost almost twice the initial estimate.
The PhD work explicitly defines and employs the relationships between different data items and their semantic descriptions. It allows a better search and reuse of already existing data and reduces the effort of project management and thus increases the probability of project success. Information of already finished projects can be consulted and be used as input to plan new projects, as well as to monitor progress and risks of projects underway. 4.2 Ontologies Ontologies are the backbone of Semantic Web technologies. They establish a common understanding of a domain by making the shared conceptualization explicit in a machine-accessible manner.
An ontology represents the domain knowledge by describing its
concepts or entities, and the relationships between these concepts
in a precise, detailed way, so that all relevant knowledge of the
domain is actually made explicitly. PROMONT [9], for example,
is a basic project management ontology which formalizes the
typical elements for project structuring. Another already existing
ontology is Harmonise [
In this PhD the innovation is to develop a project management domain ontology, as well as the development of related ontologies, like date, time or events and the matching of the ontologies. However, the primary focus will be on IT projects. 4.3 Semantic Desktop Semantic Desktop systems provide personal information management where a user can store personal information items (documents, emails …). These information items are interpreted as Semantic Web resources, identified by a Uniform Resource Identifier (URI) and all data tasks are accessible as RDF Graphs.
Ontologies allow users to qualify their information with their own
words and enable the relationships between different information
tasks (documents, contact information, calendar …). There are
already some Semantic Desktop systems available, like Gnowsis
[
Current work in the field of Semantic Desktop is to enable
collaboration between such systems [
In the PhD work one existing Semantic Desktop system, either IRIS3 or Gnowsis4, will be adapted for including project relevant information and therefore an interface between a Semantic Desktop and the PM system will be built. The idea is that project members can search on their personal desktop for project relevant information. 5. Technical solution The PhD work is still at the beginning and thus, this technical solution is a first draft how it should look like.
The ontologies will be developed in OWL (Web Ontology Language) using Protegé. The programming language of the application depends on the selected project management system, which will be evaluated. It was decided that the PhD application will rely on an OpenSource solution. The database will be either PostgreSQL or MySQL and the application will run on a Linux server.
The implementation will follow a client-server architecture. The client uses Web Services, interacting with the Project Management application. The application includes the project relevant functionalities as well as the semantic ones. The application communicates with the database, ontologies and data storage via standard interfaces. The types of these interfaces depend on the finally selected PM system. Figure 1 gives an overview of a draft architecture of the PhD project (SemProM is short for Semantic based Project Management).
The fat-client offers functionalities, such as upload and download of documents, editing (e.g. tasks, milestones, calender), annotation of information, etc. With this client in the Semantic Desktop, relationships are enabled between project management information and information on the personal desktop. 1 http://project-open.org 2 http://www.dotproject.net 3 http://www.openiris.org/ 4 http://www.gnowsis.org/ Another possibility of implementing the client is a conventional Web Browser. Therefore an evaluation is needed for studying the feasibility of useful functionalities. However it is not guaranteed that all the functionalities of the fat-client are feasible for the Web Browser. 6. Future work The work is at the very beginning and hence an in-depth study of the available literature is necessary to understand the general terms and methods which already exist. In addition, already existing ontologies as well as project management systems must be analyzed and proofed if they can be taken as basis for this PhD work.
During this study, a detailed architecture of the prototype should be designed to specify all relevant interfaces and potential difficulties as well as a time plan for the programming part, where the building of the ontologies and their combination, the extension of the project management system, the extension of the Semantic Desktop, the interfaces and the combination of all these tasks are part of it.
Furthermore use cases and requirements will be defined by means of evaluations.
Based on these results the domain ontology for project management and the other ontologies will be developed, implemented and combined.
After these steps the prototype as well as the interface between the Semantic Desktop and the PM system and the fat-client on the Semantic Desktop will be implemented and tested by the staff of the Institute of Software Technology and Interactive Systems. 7. Scientific Contribution Current research in the area of semantic desktops focuses on integrating those systems in order to interchange information. The work in this thesis concentrates on integrating a Semantic Desktop and a project management system extended by semantic functionalities. Thus, this thesis is in line with current research in this area and an exchange of information between them and us is preferable.
The main goal of this thesis is to accomplish an integration of two systems having different business justifications – a project management tool and a semantic desktop. Best to our knowledge, this has not been done before. The integration challenge is twofold: firstly, both systems must be aligned on the semantic layer. Secondly, information items on the semantic desktop (client) must be linked on the technical layer with items on the project management system (server). The user benefits from such an integrated environment by having project management items enriched with semantic descriptions readily available on the desktop. 8. Conclusion The main advantage of this PhD thesis is the higher probability of successful delivery of projects based on better information retrieval. The use of semantic technologies may lead to less expenditure of time and therefore fewer costs. Also a better storage of project relevant information is given due to for example, metadata and relationships between information items.
By the improved relationships between different information tasks connections between previous and current projects are possible and therefore a better administration of projects is feasible. Also a faster reaction to project changes is possible because experiences and problems of previous projects and their management are traceable and visible. A Multi-Agent System for Content Trading in Electronic Telecom Markets Using Multi-Attribute Auctions
Ana Petric Faculty of Electrical Engineering and
Computing, University of Zagreb
Unska 3, Zagreb, Croatia
ana.petric@fer.hr ABSTRACT The advent of the Internet and the development of the New Generation Network (NGN) has enabled, while investments in licenses and the desire to stay competitive in the future has triggered the development of value added services (VAS). Due to high market penetration, the telecommunication industry has been facing income stagnation. Consequently, it has been shifting focus to VAS in order to increase income. When forming VAS, special attention needs to be paid to the purchase of resources (e.g., transport capacity and information resources) needed for the service creation. The fact that information resources (i.e., content) are not commodities, opens the question of what is the best (i.e., efficient) mechanism that should be used for trading. As the number of participants on the B2B telecom market increases, the need for the automation of transactions carried between them is critical. The automation of transactions should lower operational costs and speed up the service provisioning process. In this paper, we try to identify stakeholders on the telecom e-market, establish their roles and relationships and find an appropriate model which captures their transactions. Finally we consider the use of multiattribute auctions for content trading in telecom markets.
Categories and Subject Descriptors I.2.11 [Distributed Artificial Intelligence]: Intelligent agents, Multiagent systems. J.4. [Social and Behavioral Sciences]: Economics. I.6.5 [Model Development]: Modeling methodologies General Terms Management, Design, Economics. 1. INTRODUCTION The advent of the Internet and the development of the New
Generation Network (NGN) provide connections which enable a
particular lifestyle that is aspiring to digital humanism where
people’s daily activities are becoming more digitalized,
convenient and intelligent [
The research problem addressed in this paper concerns the automation of business processes related to the creation of VAS that are traded on the telecommunication electronic markets (emarkets). There are two types of resources needed for the creation of telecom VAS. They are the information resources (i.e., content) the service is based on and the transport capacities needed for service provisioning. The telecom market is divided into two submarkets, the B2B (Business-to-Business) market and the B2C (Business-to-Consumer) market. Our research is focused on B2B telecom e-market trading with information resources using multiattribute auctions.
The rest of the paper is structured as follows. Section 2 describes the participants on the telecom e-market. Section 3 describes the phases we need to go through in order to conduct a transaction on the B2B telecom e-market. Section 4 addresses general auction mechanisms and presents multi-attribute auctions. Section 5 states the main questions of this research effort and proposes some answers. 2. TELECOM E-MARKETS The appearance of new stakeholders on the B2B telecom market had to be taken into account so new business models were formed. One of the long-term objectives of the NGN is to support business models that open the market to emerging service providers [14]. The volume and dynamic nature of VAS offered in the NGN place novel demands and challenges on telecom stakeholders. In this newly developed situation it is not enough just to adequately respond on the existing requests but also to intelligently anticipate the development of the future events and adapt to their environment. In order to understand the relationships between stakeholders and the way they interact it is important that their roles are well classified. We use the classification determined in [8] as shown in Figure 1.
Consumers are service users that have at their disposal various devices (e.g., mobile phone, laptop, PDA) and are connected through various access networks (e.g., 3G, WiMax). Access Provider ensures telecommunication access for service consumers. Service Provider facilitates a variety of basic and integrated services for consumers enabling easy content consumption. Carriers provide a transport service for the data traffic and they usually buy bandwidth from Network Infrastructure Owners who provide transmission lines. A large number of Carriers are at the same time also Network Infrastructure Owners. Wholesaler of Capacity provides lowercost transmission and storage capacity. Content Owner possesses the information in its original format while Content Enabler converts this information to a format eligible for the transmission over heterogeneous networks. Content Provider is at the same time Content Owner and Content Enabler. Wholesaler of Content provides lower-cost content. Server Infrastructure Owner provides storage capacity and server functionality. Information Enablers enable information resources while Transport Enablers provide transport of information resources through the various networks swiftly and seamlessly.
We are focusing on the B2B e-market since it is widely believed
that it will become the primer way of doing business [
n
USE n n n
FACILITATE n
n STORE & ENABLE n TRANSPORT
n Content Owner Server Infrastructure n
Owner Wholesaler of Content n n n
Carrier
n Network Infrastrucutre n
Owner
Wholesaler of Capacity n
3. B2B TELECOM E-MARKET
The BBT (Business-to-Business Transaction) model [15]
systematically analyses processes in B2B e-markets. The
proliferation of auctions on the Internet, and the dynamic nature
of auction interactions, argues for the development of intelligent
trading agents which act on behalf of human traders (i.e., buyers
and sellers). Intelligent trading agents can also be used to
impersonate stakeholders in the environment of the NGN in order
to enable automated interactions and business transactions on the
telecom markets [
From the BBT model perspective [15], we can formally identify six fundamental steps which must be executed in order to successfully complete one transaction in a B2B environment.
These steps are as follows (Figure 2): 1) partnership formation, 2) brokering, 3) negotiation, 4) contract formation, 5) contract fulfillment, and 6) service and evaluation. B2B negotiation is complex since it typically involves larger volumes, repeated transactions and more complicated contracts. This is the reason why most researchers have concentrated on the negotiating phase of B2B market transactions.
The partnership formation phase usually includes forming of a new virtual enterprise or finding partners to form a supply chain.
A virtual enterprise represents a form of cooperation of
independent stakeholders which combine their competencies in
order to provide a service [
Negotiation is a process which tries to reach an agreement regarding one or more resource attributes (e.g., price, quality, etc.). Each stakeholder in the negotiation process is represented by an intelligent trading agent that negotiates in his behalf (e.g., Information Agent trades in behalf of Information Enabler). The trading agent uses a negotiation strategy suitable for the type of auction applied (i.e., negotiation protocol) on the market. The negotiation protocol defines the rules of encounter between trading agents. It should ensure that the negotiation’s likely outcome satisfies certain social objectives, such as maximizing allocation efficiency (i.e., ensuring that resources are awarded to the participants who value them the most) and achieving market Procurament auctions (Multi‐attribute Auction) Double‐sided auctions (Continuous Double
Auction ‐ CDA)
NEGOTIATION
CONTRACT FORMATION
BROKERING CONTRACT FULFILLMENT
PARTNERSHIP
FORMATION BBT model SERVICE AND EVALUATION equilibrium [9]. The negotiation strategy represents a set of rules that determines the behavior of a trading agent.
The negotiation process can be either distributive or integrative
[
One party tries to minimize loss and the other party tries to
maximize gain. Distributive negotiations are also characterized as
“win-lose” negotiations. The continuous double auction (CDA),
which is suitable for transport capacity trading in a B2B e-market,
represents a distributive type of negotiation in a multi-unit auction
with multiple buyers and sellers [
In integrative negotiations, multiple issues are negotiated while the parties involved have different preferences towards these issues. For example, two information enablers may want to sell multimedia information resources to a portal provider, but one is primarily interested in the sale of news, whereas the other is interested in the sale of movie clips. These variant valuations can be exploited to find an agreement resulting in mutual gain. If their preferences are the same across multiple issues, the negotiation remains integrative until opposing interests are identified. In such a case, both parties can realize gains: consequently, another name for this class of negotiations is “win-win” negotiations. A multiattribute auction represents an integrative negotiation process which can be used for trading with information resources.
Last three phases include termination of negotiation where negotiated terms are put in a legally binding contract (contract formation), carrying out the transaction agreed in the contract (contract fulfilment) and traders evaluating the received service (service evaluation). Due to legal issues and subjective judgments it is not likely that these phases are going to be automated with the use of intelligent agents. 4. AUCTIONS Auctions, due to their well defined protocols, are suitable enablers of negotiations in e-markets. The variety and value of goods that are sold in auctions has grown to tremendous proportions.
Auctions are defined as a market institution that acts in pursuit of
a set of predefined rules in order to compute the desired economic
outcome (i.e., high allocation efficiency) of social interactions
[
Negotiation on commodities, such as transport capacities, focuses
mainly on the price of the item. These items are mostly sold in
conventional single-attribute auctions. On the other hand,
complex items such as information resources often require
negotiation of several attributes, and not just the price [
The first step in a multi-attribute auction is for the buyer to specify his preferences regarding the item he wishes to purchase.
Preferences are usually defined in the form of a scoring function
based on the buyer’s utility function [
http://www.cindywaxer.com/viewArticle.aspx?artID=149 (Business 2.0 magazine) SERVICE PROVIDER (BUYER)
Business PROVIDER AGENT (BUYER ROLE)
INFORMATION
ENABLER (SELLER #1)
Business INFORMATION AGENT
(SELLER ROLE) INFORMATION
ENABLER (SELLER #2)
Business INFORMATION AGENT
(SELLER ROLE) INFORMATION
ENABLER (SELLER #5)
Business INFORMATION AGENT
(SELLER ROLE) [11]. When trading with multimedia content there are several attributes that are negotiated on; the quality of the audio and video content (i.e., audio bit rate, resolution of the video), type of the information provided (i.e., music, video clips, games, news, sports, weather…), time of origin of the content (e.g., two days old weather forecast is of no use, one minute old stock market news could be worth a lot), reusability of the content (i.e., using the content in forming various services), potential number of users interested in this content, and the price. An example of trading with multimedia content by using multi-attribute auctions is shown in Figure 3 where several agents posing as sellers offer different multimedia (i.e., audio and video) content while the agent posing as a buyer must decide which content holds the highest utility for him and then buy the content in order to resell it further on the B2C e-market [18]. 5. PROPOSAL OF THE RESEARCH PLAN The research plan consists of four stages2. The aim of the first phase is to explore the telecom market, identify participants on the market, establish their roles and relationships, and, finally establish with what goods and services are being traded on the market. This phase is completed and is described in Section 2. We can se that a service provider actually manages a supply chain by buying information and transport capacities on the B2B telecom 2 I am a 3rd year PhD student. e-market, forming value added services from purchased goods and then selling those services to consumers on the B2C telecom e-market.
Due to the lack of research related to B2B telecom e-markets [13]
(i.e., most research is related to the B2C telecom e-market [
Since the B2B e-market includes repeated transactions with existing and/or new business partners, a new phase should be introduced into the BBT model. This phase will be in charge of collecting knowledge regarding the state of the e-market, processing information collected in the service evaluation phase, and deciding on the changes that need to be applied in the next round of negotiations.
The third phase is dedicated to the negotiation phase of the BBT
telecom model. Well defined and widely researched CDA is used
for trading with transport capacities. Consequently, we decided to
focus on multi-attribute auctions for trading with information
resources (i.e., content). In order to trade with content, the first
step is to define relevant attributes and form an ontology which
adequately represents multimedia content. The next step is to
study existing models of multi-attribute auctions using different
approaches (i.e., defining utility functions [
After studying the existing models, we plan to choose the best features from each approach and try to incorporate them into a new unified model most suitable for content trading. The new approach will be incorporated into to the multi-attribute auction mechanism based on the English auction. Due to the specifics of the B2B telecom e-market (e.g., larger values of single transactions, repeated transactions, and a smaller number of participants than on B2C e-markets) the goal is to create a balance between maximizing the allocative efficiency of the B2B market and maximizing revenue or the expected utility of bid takers characteristic for multi-attribute auctions.
The fourth phase will be devoted to implementing the multiattribute auction with agents as representatives of telecom stakeholders using the JADE (Java Agent DEvelopment Framework) agent platform and evaluating the designed mechanism with the existing mechanisms mentioned in the previous paragraph. 6. ACKNOWLEDGMENTS The work presented in this paper was carried out within research projects 036-0362027-1639 "Content Delivery and Mobility of Users and Services in New Generation Networks", supported by the Ministry of Science, Education and Sports of the Republic of Croatia, and "Agent-based Service & Telecom Operations Management", supported by Ericsson Nikola Tesla, Croatia. [8] Fischer & Lorenz (European Telecommunications
Consultants). 2000. Internet and the Future Policy Framework for Telecommunications. Report for the
European Commission [9] Friedman, D., and Rust, J. 1993. The Double Auction
Market: Institutions, Theories, and Evidence. Perseus
Publishing, Cambridge. [10] Furregoni, E., Rangone, A., Renga, F. M., and Valsecchi, M.
2007. The Mobile Digital Contents Distribution Scenario. In Proc. of the Sixth Int. Conference on the Management of Mobile Business (Toronto, Canada, July 9-11, 2007). IEEE
Computer Society, Washington, DC, USA, 32. [11] Ghys, F., Mampaey, M., Smouts, M., and Vaaraniemi, A.
2003. 3G Multimedia Network services, Accounting, and user profiles. Artech House, Inc. Norwood, MA, USA [12] Griffin, D., and Pesch, D. 2006. Service Provision for Next
Generation Mobile Communication Systems – The Telecommunications Service Exchange. IEEE Transactions on Network and Service Management. 3, 2. (Second Quarter 2006), 2-12. [13] Griffin, D., and Pesch, D. 2007. A service oriented marketplace for next generation networks. In Proceedings of the 6th international joint conference on Autonomous agents and multiagent systems (Honolulu, Hawaii, May 14-18, [14] Hanrahan, H. 2007. Network Convergence: Services,
Applications, Transport, and Operations Support. John Wiley & Sons, Inc. [15] He, M., Jennings, N.R., and Leung, H. 2003. On Agent
Mediated Electronic Commerce. IEEE Transactions on Knowledge and Data Engineering. 15, 4 (Jul.y/Aug. 2003), 985-1003.
Development of an extended selection algorithm for projects in a project portfolio
Markus Brandstätter Dr. Julius-Hahnstrasse 2/1/18
2500 Baden, Austria markus@brandstaetter.cc ABSTRACT This paper touches the current state of the art for selection algorithms of projects in a project portfolio and extends the existing approaches by prioritizing projects according to their strategic contribution based on a Balanced Scorecard (BSC). Balanced Scorecards(BSCs), project portfolios, selection algorithm 1. INTRODUCTION
Project Portfolio Management (PPM) is a field of research that becomes more and more important – mostly driven by economic thinking, competition and regulations. Especially regulations like the Sarbanes-Oxley Act (SOX) require companies to document decisions and the according decision making process. PPM represents a framework for doing this in the environment of projects and project portfolios. The paper gives a short overview on the fundamentals of PPM and focuses on the selection process for projects to become part of a portfolio. The other parts of the PPM-process are not touched in this paper – reference for continuative literature is made at the respective sections in the paper.
The algorithm developed in chapter 3 can be applied to any kind of project in any industry, as the criteria are based on Project Management Standards and on company-specific criteria coming from the implementation of a Balanced Scorecard (BSC). If the company has already a BSC in place, the algorithm can be directly implemented.
After the development of the algorithm, the paper shows the application with some test-data taken from a bank. This part focuses on optimizing a sample portfolio with test data and given constraints.
The methodology itself cannot guarantee the success of the projects in a portfolio, as this depends on various other factors as well, but it ensures the traceability of the selection of the projects in this portfolio.
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. 10th Int. Conf. on Electronic Commerce (ICEC) ’08 Innsbruck, Austria Copyright 2008 ACM 978-1-60558-075-3/08/08 ...$5.00.
STATE OF THE ART IN PROJECT PORT
FOLIO MANAGEMENT (PPM)
The role of Chief Economic Officers (CEOs) in the past was mainly driven by optimizing economies of scale for the companies they were leading. Research and development was a minor success factor as the life-cycles of products lasted over years and the competition was driven by the price rather than by unique selling points for specific products. Over the time markets evolved and products moved closer and closer to the special requirements of various customer groups – the diversification increased.
The companies needed to adapt their organizatorical structures and processes in a way to be more efficient and to react quicker to the needs of the market. In parallel legal restrictions like the Sarbanes-Oxley Act (SOX) and the right of the stakeholders to understand and follow up on the decisions made by the board forced them to increase their level of Corporate Governance. Criteria like Profitability, Return on Investment (ROI) and Windows of Opportunity were extended by topics to optimize the implementation of the company’s strategy: • What mix of potential projects will provide the best utilization of human and cash resources to maximize long-range growth and return on investment for the company? • How do projects support strategic initiatives? • How will the projects affect the value of corporate
shares (stock)? To answer these questions, the projects within the company needed to be managed following the mission and implementing the strategy of the company. This is what led to the current best-practice in the implementation of PPM.
Historical development of PPM
In 1952 Harry Markowitz described the Modern Portfolio Theory (MPT) for the first time in his seminar paper ”Portfolio Selection” in the Journal of Finance [MAR52].
In 1981 F. Warren McFarlan applied MPT to the management of projects. In the Harvard Business Review entitled ”Portfolio Approach to Information Systems” [MCF81] he recommended employing a risk-based approach to select and manage projects.
In 1994 the US Government Accountability Office’s (GAO) report ”Improving Mission Performance Through Strategic Information Management” [GAO94] described the private sector organizations using a portfolio investment process to select, control and evaluate projects.
In 1998 the GOA published the ”Executive Guide: Measuring Performance and Demonstrating Results of IT investments” [GAO98]. Portfolio management and analysis were pointed out as one of four strategic enterprise objectives.
Since the introduction of SOX in 2002, companies noted at the stock exchange have a special demand to be transparent in the use of their capital and the actions they pursue.
Project Portfolio Management proved to be one possibility to comply with the regulatory standards by effectively managing the companies’ resources. Figure 1 describes the level of implementation as of 2005 [PS05]. 2.2
Definition of PPM
As projects became more and more important over the years, traditional organizations organized around operations where extended with a second field for project execution.
They were controlled separately from each other and many of the stakeholders recognizing the shortcomings of this approach considered PPM to be the bridge between the two worlds like shown in figure 2. The reasoning behind this is based on the fact that operations and projects utilize the same resources but have different views on them. The functional departments focus on business performance, project manager focus on their projects’ performance; the satisfaction of the stockholders is of interest for the business whereas projects are more interested in the satisfaction of the stakeholders – just to give two examples. In fact PPM is much more than that; following the definition of [CEK99] and [CEK01] PPM needs • to maximize return and achieve financial goals • to maintain the competitive position of the business –
to increase sales and market share • to properly and efficiently allocate scare resources • to forge the link between projects selection and business strategy • the portfolio is the expression of strategy – it must
support the strategy • to achieve focus – not doing too many projects for the limited resources available and providing resources for the great projects • to achieve balance – the right balance between longand short-term projects and high-risk and low-risk ones, consistent with the business’s goals • to better communicate priorities within the organiza
tion vertically and horizontally • to provide better objectivity in project selection and
weed out bad projects This alters the original picture towards a new understanding of PPM: PPM acting as a hub servicing various interests and functions (figure 3).
• Strategic and tactical plans the proper prioritization of projects according to their relevance to the strategy and the progress of execution to achieve the targets set need to be monitored. • Resource Availability resources for upcoming projects need to be scheduled and planned carefully as there is normally some lead time for all of them: e.g. human resources need either to be trained or hired, financial resources like loans need to be applied for. • Budget and Cash Flow budgets for projects need to be cross-checked and the cash flows determined to plan the needed financial resources. • Scope, Change and Cost Control the scope of any of the projects in the portfolio needs to be monitored tightly as all the dependencies to other projects in the portfolio depend on it. Changes might affect not only a project but the whole portfolio. • Opportunity Management if there are opportunities for optimizing the portfolio, they need to be recognized and managed. This function has a wide field of activity starting from the recognition of newly raised dependencies between projects in the portfolio over changes in the market to better allocate resources to moves of competitors that might change the implementation plans for the company’s strategy. • Demand (Internal Projects) besides the strategic projects of a company there is also the need for projects that might not have direct impact on it but are required to improve certain processes. • Project Control and Performance the ongoing projects need to be monitored in terms of classical project management procedures to understand the progress and realize the impact of the evolvement on other projects in the portfolio. • Resource Allocation the resources need to be distributed over the projects to optimize the possible output. As the resources are limited this needs to be handled in with a portfolio optimization approach like the Modern Portfolio Theory (MPT) from H. Markowitz. • Risk Assessment and Management the risks for the projects and the whole portfolio need to be accessed during the selection of the projects and during the whole life-time as risks evolve over time.
The responsibility of PPM is to protect the company from unexpected risk. • Business Performance the execution of the projects and the portfolio is very important but the crucial point for the company’s success is, if the projects delivered return the business value expected. Therefore the implemented projects need to be reviewed after their closure for the return of the investment and the conclusions for ongoing and future projects need to be drawn. 2.3
The Process Model of PPM
In contradiction to the execution of projects PPM does not have a defined beginning nor has a defined end as it is an ongoing process. However the process can be divided into five phases which are often separated with methods like the Stage Gate R Process, meaning that they are only allowed to enter the next phase if the phase before is finished and certain criteria are fulfilled. The phases for PPM are defined as follows: • Identification of needs, goals and objectives in the first step the requirements for the portfolio are defined. The needs describe the reasons why the company implements PPM. The goals and objectives define the targets to achieve and the measures to quantify them. Taking these as a baseline the selection criteria are built or updated to choose the proper projects for the portfolio. As expectations towards PPM evolve over time and the acceptance and success of PPM depends on clear expectation management, this step is not defined as a one-time preparation step but is an integral part of the PPM-life cycle. Important to mention is that the objectives should stay as stable as possible over the time – a change in the objectives for the PPM means that the traceability of the portfolio is disturbed. • Selection of the best combinations of projects (the portfolios) the quality of the portfolio depends to a large extend on the quality of the selection-criteria defined in the first step. In addition the current business strategy and actual targets are taken into consideration to pursue the right set of projects. Details will be discussed in the chapter dealing with criteria for selecting projects. • Planning and execution of the projects this step deals with the scheduling and the conduction of projects in the portfolio. It must not be mixed up with the planning and execution of projects in terms of project management as the focus on this phase is on the state of the portfolio and the contribution of each single project to the portfolio and not on the details of the projects in the portfolio. • Monitoring portfolio performance the monitoring focuses on the delivery of the expected project deliverables and their contribution to the development of the portfolio. Variations of the plans are detected and corrective actions or change requests are set. The main deliverable of this phase for the board of the company is a dashboard providing them with the information what the actual status on the implementation of the strategy is. • Realization of benefits the last phase in the cycle is used to compare the implementation of the project deliverables and their business impact to the expected results. This step does not only provide information on the achievements of goals but needs to be used to question the reasons in case of failure as well. They might give new or additional input to the first cycle again and help defining the needs, goals and objectives.
There is no general recommendation on the duration per iteration of the life cycle as a reasonable time frame depends on the projects in the portfolio. However, for most of the portfolios a benchmark should be a month.
The portfolio life-cycle described up to now leaves open, how the Project Management methodology is embedded. The best-practice solution which is proven by various implementations is shown in figure 5.
As projects have different start-dates, milestones and enddates, they cannot be synchronized in a way that they fit in a phased approach were all of them sharing the same rhythm. Therefore PPM needs to adapt and be so flexible to handle projects in various stages of their life-cycle and still fulfil its function. The approach in figure 5 shows that ideas and opportunities are collected in the very beginning but are treated outside the PPM-Cycle itself (but in the PPMresponsibility as displayed in figure 3). After an opportunity or idea has been selected, the initiation phase for the project starts from where it continues through the whole project management life cycle (Initiation – Planning – Executing – Controlling) besides Closing. During all the time PPM oversees the project and monitors it. As soon as it comes to Closing, the project quits the life cycle, as this part is administrative only and does not impact the value delivered to the portfolio any more.
As this paper focuses on extending the existing selection algorithms, it will only deal with optimizing the value of the portfolio by using the proper set of criteria to prioritize projects. For the other process steps reference is made to the explanations in [LEV05], [PS05], [COO05], [PMI06] and [RMW07].
The existing approach introduced in chapter 2.3 contains weaknesses in terms that some major aspects of an efficient PPM are not fulfilled: 1. The existing project portfolio optimization models are based on the methodology of Markowitz [MAR52] but do not consider one important fact: Markowitz based his theory on financial portfolios. The difference between financial and project portfolios in this relation is, that financial ones are continuously distributed whereas project ones are discretely distributed. This is because in financial portfolios shares or derivatives can be sold or bought in arbitrary pieces whereas projects can be executed or not – the execution of 50% of a project by gaining 50% of the benefit is unrealistic. In projects the earned value cummulates over all deliverables of the project and the benefit cannot be split or partially fulfilled by a certain number of deliverables. Explanations on the difference between continuous and discrete distribution can be found in [KRE98]. 2. Further the projects are evaluated by themselves but not by their value contribution they have with their dependencies to other projects. This means that projects with a low value by themselves but acting as an enabler for high-valued projects might not be implemented. A special case in this coherence are projects in a portfolio that do not provide a business need itself but are obligatory (e.g. needed to fulfil regulations set by the government). 3. The quantifying measures to prioritize the projects are related to financial figures only. They do not take the influences on other key performance indices (KPIs) relevant for the company into consideration. Therefore this approaching is lacking to optimally support the strategy of the company. 4. The assumption for most existing optimization algorithms is, that the limitation of human resources can be resolved by investing additional money to buy additional ”Know-How”. In reality, this is normally not true as sensitive and important projects require special people with high sophisticated skills.
The approach that should be developed in this chapter tries to address all these issues and will provide suggestions to resolve them in a way that project portfolios can be optimized fully considering them.
The distribution of a project portfolio
The first weakness identified goes along with the distribution of a project portfolio. [GRU05] explains that
”The Efficient Frontier curve shows all of the best possible combinations of project portfolios and the value that can be created with available capital resources in an unconstrained mode.” and further
”The Efficient Frontier shows the opportunity cost of investing an additional dollar versus the additional value received.” The second statement implies that an arbitrarily chosen amount of money adds additional value to the project defined by a certain function. This would also mean that projects can be split in smaller pieces by delivering a smaller value that can be determined.
In reality this does not work out. Imagine a car manufacturer that needs to develop two new cars: the first one takes development cost of 500 million dollar and the second one of 600 million dollar, the budget of the company is 800 million dollar. Following the Efficient Frontier approach would mean that the company could e.g. run the project for the first car and invest the remaining money into the project for the second car. Obviously there is a value for project one if we assume that it is finished successfully and it goes into production and into sales – the money invested into the second project does not provide any value so far: a car where the product development is not finished can neither be produced nor sold.
This needs to be repeated for possibilities of k (the number of projects that can be executed in parallel). In theory k can be any number between 1 and n because without dependencies and limitations all projects could be executed; limiting k only makes sense if the stakeholders do not want to support more than a maximum of k projects at the same time: n X k! (n − k)!
The first complexity to be added are the dependencies.
Formally it means that for a portfolio at t0 only projects that to not rely on any other project can be executed. All other projects need to wait for the finishing of their predecessors; this reduces the complexity of the project portfolio by or n X k! (m − k)!
(n − m)! k! (n − m − k)! where m are the number of dependent projects in the portfolio. This finding needs to be handled with high caution as it might lead to a failure: prioritizing now the portfolio based on the value of the projects in t0 would lack the vision that is necessary in PPM: a future oriented approach should keep in mind all combinations of projects including the value of each path – a sample path diagram based on dependencies in a project portfolio is shown in figure 6.
Thus the algorithm needs to be extended to not only look at t0 but considering the whole timeline until the finalization of the last project to optimize the portfolio on the maximum expected benefit out of all options in the future. Therefore every project needs to be listed with all its dependencies: (1)
The issue can only be solved by changing the approach from arbitrarily changes in size to changes in terms of full projects – this implies further that the type of distribution that needs to be used is not a continuous one like [MAR52] used for financial portfolios but a discrete one: a distribution that shows all possible portfolio combinations. For simplification purposes at the beginning the following topics are not considered – they will be added later on: • dependencies to other projects • observations beyond the point in time t0 • obligatory flags for projects • limiting constraints This determines the number of portfolios alternatives to be a combination of n different projects taken k at a time, without repetitions or n k k! (n − k)!
P6 P2 P5 - P7
P3 Project
- P8
The value of project options in a project
As the various paths in the project portfolio are known based on their dependencies, the next step is to benchmark every option for its value. Naturally the prioritization would take place by sorting them by a certain selection criteria.
The issue using this approach is, that the duration of the project(s) is not considered and therefore the return on the various options is not evaluated on the same baseline. For this reason the indicator needs to be discounted over the duration of the project respectively the duration of projects within the option:
Option Value = X Indicator p (1 + r)dp where p represents all projects in the option, r is the discount rate per period and dp is the period from the beginning of the portfolio’s perspective to the end of the considered project (not the portfolio!). Caution needs to be taken in case the indicator chosen is already discounted like the NPV or a derived one. An effective indicator for the Option Value will be introduced in section 3.4.
There might be situations where projects in a portfolio are obligatory e.g. for regulatory reasons. These projects might not return any direct value to the company. Therefore they need to be incorporated separately as they would otherwise never make it into the project portfolio.
The solution is to mark them and all the projects they depend on directly and indirectly as mandatory, considering them before the priority list given by the indicator.
By now the approach addresses the issues one and two identified at the beginning of the chapter – the next step must be to find a solution to problem number three: the solution described so far relies on financial KPIs only, but does not consider further influences of the project on the 3.3
Quantifying measures beyond financial KPIs
This section deals with the fact, that indicators cannot only be taken from the financial information that goes along with the execution or finalization of the project but also with the influence on and from other components of the success of a company. This paper distinguishes between two types of such indicators: the ones derived from a Balanced Scorecard and others taken from standard project management methodology defined by the Project Management Institute (PMI) respectively well known indicators out of standard project management. 3.3.1
Measures from a Balanced Scorecard
The first possibility to extend the traditional view is to follow a Balanced Scorecard (BSC) [KN96] approach by taking the measures and KPIs identified in a BSC to determine the influence of a project on this BSC. This implies that the projects can be evaluated on their contribution to the strategy that is defined in the BSC.
The BSC identifies objectives and the influences between these and tries to bring them down to factors that do not represent aggregates figures but are pure values that cannot be further decomposed (so called α-figures). Their transformation to the operational KPIs is defined in the mathematical model of the BSC which provides the first of the transformations needed to get the target values for prioritization in this model. The figure below describes the projects in a portfolio (P1 . . . Pm), the BSC input variables (Bα1 . . . Bαn) and the BSC output variables (Bβ1 . . . Bβn). The B stands for BSC – there will be additional input- and output-variables described afterwards, so this identifier is needed.
BSC Input Variables
Bα1 . . . Bαn
For the evaluation of projects it is important to understand how the finished project will change the α-figures of the BSC. Out of the transformation (a n:m transformation between input- and output-variables) the expected change in the strategic figures can be calculated (the β figures).
Measures taken from Project Management
So far this section dealt only with KPIs determining the alignment of a project with the strategy of the company.
In addition there are other KPIs that deal with project inherent data and are necessary for the selection process as well. The ones the paper is referring to are the ones of the Project Management Institute (PMI) defined in the Project Management Body of Knowledge (PMBOK R ) [PMI04].
The input parameters (or α-figures) can be defined as follows: • Rate of the skill describes the rate to be paid for the specific skill. The scale chosen needs to be the same as the scale the demand is given in and needs to be available or estimated for all periods the project is planned to be executed in. • Investments
depict the investments planned within the project. • Investment cost determine the cost that go together with the investments described. • Risks the risks that go along with the project need to be identified. • Lowest possible impact for every risk identified • Probable impact for every risk identified • Highest possible impact for every risk identified • Direct dependencies to other projects the dependencies included may only be mandatory dependencies for the execution of the project. Sometimes they become mixed up with so called discretionary dependencies sourcing from e.g. resource shortages – they need to be filtered and removed as the selection algorithm would not work efficiently in this case. Section 3.5 shows that this kind of dependencies comes from constraints within a portfolio.
Out of these factors the following output-parameters (or β-figures) can be derived. Formally they underlie the same kind of transformation that can be seen with the factors from the BSC – projects in a portfolio (P1 . . . Pm), the project input variables (Pα1 . . . Pαn) and the project output variables (Pβ1 . . . Pβn).
Project
P1 . .
.
Pm Project
P1 . .
.
Pm
P1Pα1 . .
.
PmPα1
P1Pβ1 . .
.
PmPβ1 Project Input Variables Pα1 . . . Pαp Project Output Variables Pβ1 . . . Pβq P1Pβq . .
.
PmPβq Total Labour Cost (TLB).
The cost of labour depends on the demand for specific skills and their rate. This formula is only to calculate the cost of labour – at this point in time it is not yet considered that the availability might be an issue; it will be discussed later on during the further development of the algorithm.
TLB =
X X rateps × demandps p s where p represent the periods of the project and s the skills needed.
Total Investment Cost(TIB).
The investments planned within the project – also important to calculate depreciation for the spendings on inventory goods out of a project, which can also be used as an indicator in the prioritization of the portfolio (e.g. percentage of the project budget that can be activated for depreciation):
TIB =
X X investmentpi p i where p represent the periods of the project and i the investment needed.
Total Risk Cost (TRB).
Every risk in the project needs to be quantified in a way that the monetary value that goes along with it becomes determined. Therefore the lowest possible impact, the probable impact and the highest possible impact are estimated and weighted for every risk:
TRB = r X x × lir + y × pir + z × hir
x + y + z where r represents the risks in the project, li the lowest possible impact, pi the probable impact, hi the highest possible impact and x, y and z the weights. Further explanations on the estimation and calculation of risk can be found in [BRA07].
Total Project Budget (TPB).
The three figures discussed summarize to the Total Project Budget.
TPB = TLB + TIB + TRB (6) or
TPB p s p i r X X rateps × demandps X X investmentpi X x × lir + y × pir + z × lir
x + y + z Planned Value (PV).
This indicator is the baseline for Earned Value Methodology (see also [PMI04] p. 172–176 and [PMI05]) and the application of all budget related control mechanisms in a project. It is similar to the TPB but does not contain the risk budget. The reason behind is, that the PV is the basis all efforts within the project are tracked against – if risk cost would be included in this figure, non-occurred risks would be counted as success to manage the project below budget.
Further the point in time for a possible incident cannot be determined a priori and therefore a valid cost plan could not be provided.
PV = TLB + TIB
(9) p s p i PV
X X rateps × demandps
X X investmentpi or Total Effort (TE).
The total effort represents the timely effort invested in a project and is normally measured in man-years.
TE =
X X demandps p s
(10) where p represent the periods of the project and s the skills needed.
Building the quantification criteria
Obviously all of the factors determined (PmBβ1, . . ., PmBβo; PmPβ1, . . ., PmPβq) need to be used to prioritize a portfolio effectively. This introduces two new problems: • a standardization of the β-figures is needed, as most of them have different measures and scales. This is close to impossible because how should e.g. ”Customer Satisfaction” and ”Education days of an employee” be measured on the same – still meaningful – scale? • a weight for every β-figure needs to be calculated to be in the position to aggregate the factors to a significant indicator. The word ”significance” implies already that the weights need to be derived from the attitude of the decision-makers. As there is more than one decisionmaker in a team, a compromise would need to be made which is again a sub-optimal solution.
The problem can be solved by looking at the different β-figures neither considering their measurement nor their weights but still offering a transparent and comparable figure. The solution is in the calculation of the area that is spanned by the different relative β-figures in a a radar-chart (also called spider-chart) in figure 7.
The table for the base values looks as follows: Project 1 abs.
Project 2 abs.
Project 1 rel.
Project 2 rel.
1 50 120 42% 100%
2 200 40 100% 20%
Criteria 3 80 160 50% 100%
4 150 130 100% 87%
5 100 140 71% 100% Using this type of representation has several advantages: • Every figure can be presented using its measure – the only topic of importance is, that the scale is used in a way that the better the result is, the larger the distance to the zero-point of the graph needs to be. • It can be used for an arbitrary number of criteria larger
or equal than three. • The change of the scale does not change the result as
all projects are measured against the same baseline.
In the next step the area needs to be calculated for each of the sample projects. To do so, the formula for calculating the area of a polygon is given by (see also [BOU98]):
Area =
N−1 1 X (xiyi+1 − xi+1yi) 2 i=0 where N represents the number of edges in the polygon and x and y their coordinates. The last coordinate must be identical with the first one to close the area of the polygon.
To do the calculation with the items out of a radar chart, the data points need to be transformed into a two-dimensional co-ordinate system. In the first step x- and y-values of the data points are calculated taking the centre of the polygon to be the zero-point of the grid. This can be derived using trigonometric functions. Assuming – like shown in figure 7 – the line for criteria one is vertically aligned (what means 90 degree or π/2) the formula is defined as follows; let xi and yi be the x and y coordinates relative to the centre of the radar-chart for every relative value χi of the corresponding criteria βi in the radar-chart where N is the total number of criteria: xi = cos yi = sin π 2 − 2π (i − 1)
N π 2 − 2π (i − 1)
N The formula is derived the following way: xi = cos = cos = cos = cos
360 90 − N 90 −
× (i − 1) 360 × (i − 1)
N 180π 360 − 720π × (i − 1)
360N π 2 − 2π (i − 1)
N × χi × χi × χi
2π × 360
2π × 360 × χi × χi × χi
The deduction is analogical for yi. For the figures given in table 3.4, this gives the following coordinates and the area the projects cover:
P1 P2 -0,59 -0,81 -0,51 -0,70 -0,68 0,22 -0,95 0,31
The result shows, what is expected when looking at figure 7: project 2 covers a larger area and has therefore the higher value compared to Project 1 in terms of measures that are influenced by it. As discussed already, this indicator can easily combined with the formula defined in (4) to calculate the value of an portfolio option based on all the projects contained.
Constraints within a project or a project portfolio
The last remaining issue not being addressed so far is the one of constraints within a project or a project portfolio.
As a matter of fact limitations constrict the possibilities of projects to choose for a portfolio. Recent approaches try to formulate every constraint as a financial one arguing that anything else can be removed by monetary investments. In reality this is not the case as it was explained already in the description of weaknesses at the beginning of chapter 3 on the example of skills of human resources.
As discussed in section 3.3, all relevant indicators for the selection of projects are represented in the α- and the derived β-figures. This implies that also the relevant constraints for the portfolio can only hit one of these figures.
First of all, all the αs and βs from the projects and all their totals in case of combinations that could be started in t0 based on their dependencies are summarized in a matrix together with their prioritization and constraints. The order of the projects is based on the total option value of the project (except for mandatory projects) summarizing all discounted option values it is the first project in. At the bottom of the matrix, all constraints for the indicators are filled in. Additionally every constraint needs to be marked, if the constraint must not be undercut (a minimum-constraint) or must not be exceeded (a maximum-constraint):
P1 . .
.
Pm Constraint
Bα1 . . .
Input . . . . . . . . .
BSC Bαn Bβ1 . . .
. . .
Output . . . . . . . . .
Bβo . .
.
The project-specific αs and βs are not displayed in this example for space reasons. Normally the matrix is extended at their right border by the project-specific αs and βs.
All totals of αs and βs need to compared with their respective constraints. For all of them which are violated, so called ”discretionary” dependencies need to be added in the following way: the project with the lowest total option value is taken away from the portfolio of t0 and given a dependency to the project finishing the earliest after the prioritization.
This is repeated until all constraints can be fulfilled. If this is impossible (so in the worst case, the project with the highest total option value cannot be executed) the topmost project causing the conflict is removed and the procedure is restarted with all the other projects. If this extended procedure does not direct to a meaningful portfolio, the constraints are too narrow to allow a prioritization. In this case, focus need to be set on widening the constraints.
The Final Portfolio
The portfolio developed is the one that contributes best to the strategic targets of the company under the given conditions. However, the prioritization itself is not a guarantor that the targets set for the projects are also achieved. For controlling the projects in a way to have tight control on the progress, there are methods available but they are outside of the scope of this paper – a detailed description can be found at [PMI05]. 4.1
Initial situation
This chapter deals with the exemplary implementation of the approach developed. The sample setup consists out of five projects taken out of a project portfolio of a bank: 1 2 3 4 5
Name Data Warehousing (DWH) Management Information System (MIS) Customer Relationship Management (CRM) Application Processing System (APS) for loans
Collection System (CS) for overdue loans
To fulfil the quantification requirements defined in section 3.3 the model needs to rely on a BSC developed for this company and on the respective input parameters to this BSC.
The success factors defined for this sample BSC can be seen in figure 8.
The Cause-Effect model for this sample BSC is shown in figure 9. The detailed aggregation algorithms from the αfigures up to the calculation of the influence of the success factors is not discussed here in detail, as it is part of a BSC and for the algorithm in this example, only the input-figures and the output-figures of the BSC are of importance. 4.2 The distribution of the set of projects in
cluding their dependencies
Before the paper goes into detail on the α- and β-figures for this set of projects, the dependencies for this constellation are discussed. Following the formula given in (2), the complexity of five projects and their combinations give 31 possibilities to structure the portfolio in t0: P1P2, P1P3, P1P4, P1P5, P2P3, P2P4, P2P5, P3P4, P3P5, P4P5 P1P2P3, P1P2P4, P1P2P5, P1P3P4, P1P3P5, P1P4P5, P2P3P4, P2P3P5, P2P4P5, P3P4P5 P1P2P3P4, P1P2P3P5, P1P2P4P5, P1P3P4P5, P2P3P4P5
P1P2P3P4P5
In this example the MIS and the CRM project rely on the implementation of the DWH project; both systems are analytical ones and depend on various data loaded from different source systems. The APS project and Collection System project are independent from the DWH-project but the APS project is the mandatory predecessor for the Collection System (the bank could not collect overdue loans they do not have the data for). This information gives the following dependency map:
MIS
CRM DWH APS The dependency matrix for the projects is as follows:
Keeping the dependencies in mind the possible complexity of the portfolio reduces alreay from 31 to three possibilities in t0 as defined in (3):
Combination
Possibilities 2 1 2 2
In our example the contribution to the input factors (already derived from α-figures) by the projects for the BSC were identified like this: ables Absenteeism Duration of NR Applications Number sources of for
NR Applications Efficiency of NR
of Retail of for Retail Applications Efficiency of Retail Collection Degree of automatization of retail processes Number of Marketing Activities tomer Customer Satisfaction Rating Percentage of offers/deals Non-Retail tomers lost
cus
Customers tomer Retail lost ity IT Investments Building ments OutVari
TarBSC put ables ented gets Person OriProcess Excellence Market Position Gaining and taining Customers Increase Revenue Decrease Cost
ReCombination
BSC Input Vari-10 -2 -20 +1 +1 +5 +500
CS +2 +7 -5 +500 -2 -5 -2 -20 +25 +1 +100k +100 +200
-5 -1 -10 +30 +100 +2 -10 +3 -1.000 +30.000 +1 +10% Products per cus
+3 +10% +100k +150k zation input:
The corresponding output figures (β-figures) for the BSC have been calculated and bring the results for the prioriti
DWH +16 +28 +25 +500
MIS +1 +5 +500 +100
CRM +2 +7 +430 +610
APS +10 +35 +500 +200
In table 4.3 it needs to be especially noted that the operational target of ”Decrease Cost” has negative values as projects generate costs and therefore cannot contribute reducing their costs by themselves. transformation to the output variables:
So far the BSC input- and output variables have been discussed.
What is missing from the KPI point of view are the figures coming from the project input variables and their
Input Rate: Skill 1 . . .
Rate: Skill s Demand: Skill 1 . . .
Demand: Skill s Investment Cost Probability: Risk 1 . . .
Probability: Risk r Impact: Risk 1 . . .
Impact: Risk r Duration (in months) Obligatory Project Direct Dependencies Operations Cost for 3 years Project Input Variables Project Budget thereof Total Labour Cost thereof Total Investment Cost thereof - Total Risk Budget Duration (months) Obligatory Project with dependencies Full Dependencies Return on Investment (ROI)
DWH 600 . .
. 1.500 400 . .
.
400 100.000 20% . .
.
30% . .
. 1.500 600 . .
.
600 150.000 15% . .
.
10% . .
. 1.500 200 . .
.
200 150.000 40% . .
. 15% . .
. 8.000
18
CS . .
. 1.500 100 . .
.
100 150.000 30% . .
. 10% . .
. 38.000 12 4
CS 128% 188% 145% 306%
418% output matrix from the BSC:
If those input figures become calculated by the formulas explained in chapter 3.3.2 the following output matrix can be determined – also for prioritization purposes, like the 4.4
Building the quantification criteria
Of course the project budgets presented in this figure are equal with the negative decrease of cost in table 4.3. The χ-figures derived from the tables 4.3 and 4.3 are now used to build the radar-chart in figure 10. For the simplification of illustration not all criteria have been considered. The table with the base values looks as follows: for the covered area:
DWH rel.
MIS rel.
CRM rel.
APS rel.
CS rel. 100% 6% 13% 63% 9% 78% 14% 20% 100% 20% 5% 100% 86% 100% 100% 82% 16% 100% 33% -1% 77% 86% 100% 85% 73% 30% 41% 28% 68% 100%
ROI 31% 45% 35% 73% 100% This data results in the following co-ordinates and values APS CRM xy DWH xy CS MIS x y
PE 0,78 0,62 0,15 0,12 0,61 0,49 0,16 0,12 0,11 0,09
MP 0,97 -0,22 0,84 -0,19 0,05 -0,01 0,97 -0,22 0,97 -0,22
GaRC 0,14 -0,30 0,43 -0,90 0,36 -0,74 0,00 0,01 0,07 -0,15
IR -0,37 -0,77 -0,43 -0,90 -0,33 -0,69 -0,32 -0,66 -0,37 -0,78
DC -0,67 -0,15 -0,27 -0,06 -0,29 -0,07 -0,97 -0,22 -0,40 -0,09
ROI -0,57 0,46 -0,27 0,22 -0,24 0,19 -0,78 0,62 -0,35 0,28
Area 1,48 0,97 0,83 0,79 0,40
Looking at the project itself without considering the possible portfolio options would clearly favour the APS project (1,48) compared to the DWH project (0,83). As described before the project on its own is not the driving factor – it is the option value of the different options that is important.
Using equation (4), the results from table 4.4 and a discount rate of 5% shows the following results for the available options:
Option DWH DWH - MIS DWH - CRM APS APS - CS 0,79 1,16 1,67 1,42 2,13 Taking the projects being marked mandatory into the picture as well (see table 4.3) shows that the DWH-project needs to be executed for the MIS project having a obligatory status although the option DWH-MIS has the second-lowest value. Finally the following prioritization would be set:
Option DWH - MIS APS - CS DWH - CRM APS
Mandatory 1,16 2,13 1,67 1,42
The DWH-project is not given separately in table 4.4 as it is executed anyway because of the dependency.
In the next and also the last step the constraints need to be considered. Therefore a table is created as described in chapter 3.5. For the reason of clarity, the criteria already used in figure 10 and table 4.4 have been reused – the only difference to be noted is, that the ROI, the Increase in Revenue and Decrease of cost are removed but therefore an α-figure from the Project-αs is added: the demanded availability of a business analyst for the respective project measured in person days (PDs):
Project DWH-MIS APS-CS DWH-CRM APS
Total POT Option Value 1,16 2,13 1,67 1,42 +17 +12 +18 +10 +57
Min (+25)
Total Constraints
PE
MP
GaRC
BA PDs +33 +42 +35 +35 +145
Min (+60)
+525 +1.000 +455 +500 +2.480
Min (+1.000) +600 +195 +1.110
+200 +2.105
Min (+700) +550 +300 +500 +200 +1.550
Max (+700)
The table shows an obvious conflict with the person days for the business analysts needed (BA PDs). Following the procedure described in section 3.5, the options need to be eliminated buttom-up following their total option values. If this is done in this portfolio, it ends up with the following status: +33 +525 +600 +33 +525 +600
Min Min Min (+60) (+1.000) (+700) +550 +550
Max (+700)
The current status shows that the issue with the BA PDs could be solved but turned the project into conflict with lots of other constraints. Obviously the portfolio cannot be structured in a way that stay within the boundaries set.
This leaves two options: the first one is, to take the portfolio above also implying that the stakeholders need to adapt the constraints given. The second option would be to include another project to optimize the number of limits being fulfilled and focus on adapting other limits:
PE
MP
GaRC
BA PDs DWH-MIS
Total Option Value 1,16
Total
Constraints DWH-MIS APS
Total Option Value 1,16 1,42
Total Constraints +17 +10 +27
Min (+25) +33 +525 +600 +35 +500 +200 +68 +1.025 +800
Min Min Min (+60) (+1.000) (+700) +550 +200 +750
Max (+700)
In the second option the constraint of the BA PDs is violated again with a very small backlog, which might be resolved. Therefore the other constraints could be kept and the portfolio could be adjusted in the best possible way.
Most probably the company could resolve the BA PD issue and would go for the portfolio given in option 2.
The discussion in this paper showed that there are lots of improvements possible to extend the existing selection algorithms in a way to make them implementing the strategy of a company. If a company went already through the painful process of creating and implementing a BSC and is living the life-cycle process that goes along with it, the presented algorithm for the selection of appropriate portfolios is a spinoff product of the BSC and PPM. Naturally, the selection algorithm is only one part of various steps to successfully implement the strategy. Others, like the carefully planning and controlling of a project portfolio or the sustainable implementation of the project content are others that need to be dealt with seriously. Possible solutions in these fields are the Earned Value Methodology (EVM) for controlling the process or classical mechanisms for mid-term planing to compare the expected results from PPM with the realized benefits.
The challenge in the presented approach is definitely the quality of the BSC, the portfolio selection algorithm is based on. If the strategy is not described properly or the controlled measures are not the right ones to successfully achieve the vision of the company, the selected portfolio will fail the same way as the BSC will. Therefore the success of the implementation of this algorithm will heavily rely on the time that was spend for defining the strategy. This is also a lessons learned that should be taken away when project portfolios should be aligned with the strategy: the project portfolio can only be as good as the underlying strategy is.
The further steps for the PhD thesis will be the extension of the existing project portfolio life-cycle not only by the selection but also for the planning and monitoring phases. The target is to present a framework where the whole life-cycle is linked to the implementation of the strategy using BSCs.
Further the scope is exclusively to focus on optimizing the project portfolio into this direction – it is true that projects that cannot be evaluated against their benefits but might deliver unexpectedly high results will never be selected with this methodology.
For the proof of concept (POC) data will be taken from an internationally acting bank and their project portfolio.
REFERENCES
Markowitz, H. M.: Portfolio Selection. Journal of Finance, Vol. 7, Iss. 1 (1952) 77–91 F. Warren McFarlan: Portfolio Approach to Information Systems. Harvard Business Review September 01 (1981) Thomas L. Saaty: Multicriteria Decision Making: The Analytic Hierarchy Process. AHP Series Vol. 1 extended edition (1990) RWS Publications US Government Accountability Office’s (GAO): Improving Mission Performance Through Strategic Information Management. US Government Accountability Office’s (GAO) report (1994), http://www.gao.gov/special.pubs/ai94115.pdf, Viewed: 10.03.2007 Robert S. Kaplan, David P. Norton: Translating Strategy Into Action – The Balanced Scorecard (1996) Paul Bourke: Calculating the area and centroid of a polygon, http://local.wasp.uwa.edu.au/ pbourke/geometry/ polyarea/, Viewed: 09.02.2008 US Government Accountability Office’s (GAO): Executive Guide: Measuring Performance and Demonstrating Results of IT investments. US Government Accountability Office’s (GAO) report (1998), http://www.gao.gov/special.pubs/ai98089.pdf, Viewed: 20.02.2007 Erwin Kreyszig: Statistische Methoden und ihre Anwendung, 7. Auflage (1998) R.G. Cooper, S.J. Edgett, E.J. Kleinschmidt: New Product Portfolio Management: Practices and Performance. Journal of Product Innovation Management (1999) 16, 333–351 R.G. Cooper, S.J. Edgett, E.J. Kleinschmidt: Portfolio Management for New Product Development: Results of an Industry Practices Study. R&D Management (2001), 31, 361–380 Harold Kerzner: Project Management: A Systems Approach to Planning, Scheduling, and Controlling, 8th Ed. (2003) Mark Jeffery, Ingmar Leliveld: Best practices in IT Portfolio Management. MITSloan Management Review Spring 2004 Vol. 45. No. 3.
Project Management Institute: A Guide to the Project Management Body of Knowledge – Third Edition(2004) R.M. Wideman: A Management Framework for Project, Program and Portfolio Integration, Trafford Publishing (2004) Robert G. Cooper: Portfolio Management for Product Innovation. Project Portfolio Management: A Practical Guide to Selecting Projects, Managing Portfolios and Maximizing Benefits (2005) Mike Gruia: The Efficient Frontier Technieque for Analyzing Project Portfolio Management. Project Portfolio Management: A Practical Guide to Selecting Projects, Managing Portfolios and Maximizing Benefits (2005) Harvey A. Levine: Portfolio Management: A Practical Guide to Selecting Projects, Managing Portfolios and Maximizing Benefits (2005) Project Management Institute: Practice Standard for Earned Value Management(2005) James S. Pennypacker, Patrick Sepate: Integrating Project Portfolio Management with Project Management Practices to Deliver Competitive Advantage. Project Portfolio Management: A Practical Guide to Selecting Projects, Managing Portfolios and Maximizing Benefits (2005) Ray Trotta, Christopher Gardner: How to Determine the Value of a Project. Project Portfolio Management: A Practical Guide to Selecting Projects, Managing Portfolios and Maximizing Benefits (2005) Project Management Institute: The Standard for Portfolio Management (2006)