=Paper=
{{Paper
|id=None
|storemode=property
|title=Towards an Open and Scientific Approach to Innovation Processes
|pdfUrl=https://ceur-ws.org/Vol-864/paper_2.pdf
|volume=Vol-864
|dblpUrl=https://dblp.org/rec/conf/caise/DiamantiniPS12
}}
==Towards an Open and Scientific Approach to Innovation Processes==
https://ceur-ws.org/Vol-864/paper_2.pdf
Towards an Open and S ienti Approa h to
Innovation Pro esses
Claudia Diamantini, Domeni o Potena and Emanuele Storti
Dipartimento di Ingegneria dell'Informazione,
Università Polite ni a delle Mar he - via Bre e Bian he, 60131 An ona, Italy
{ .diamantini,d.potena,e.storti}univpm.it
Abstra t. Being the modern e onomy onstantly hanging and evolv-
ing, organizations are asked to develop a more exible, open and ol-
laborative mindset. In parti ular, an attitude towards ontinuous prod-
u t/pro ess innovation is seen as one of the potential solutions apable
to ee tively address the dynamism of the market. However, Business
Innovation (BI) still la ks methodologies and best pra ti es apable to
ee tively drive business users from an innovative idea to its realization
and evaluation. This work investigates the possibility to adopt a prag-
mati and systemati approa h to support business users in the manage-
ment of an innovation pro ess, with the aim to in rease the ontrol over
the pro ess and redu e the risks of failure 1 .
1 Introdu tion
Profound hanges in e onomy, so iety and te hnology are nowadays dramati ally
reshaping the environment in whi h ompanies, nations and people are used to
live. Moreover, in last years a more open so iety and e onomy is ontributing to
tear down ommer ial barriers, allowing more highly ompetitive businesses to
join the global market. Being the modern e onomy ontinuously hanging and
evolving, organizations are asked to develop a more exible, open and ollabo-
rative mindset. In parti ular, innovation is seen as one of the potential solutions
apable to ee tively address su h hallenges. Anyway, it is widely re ognized
that Business Innovation (BI) is also a risky pro ess whose out omes are often
unpredi table, ae ted by multiple internal and external variable onditions,
many of whi h are non-observable and therefore annot be properly kept un-
der ontrol. For su h reasons, design and management of innovation pro esses,
espe ially in highly ollaborative Virtual Enterprises (VE) environments, are
hallenging tasks. This is the main reason why, in fa t, there is still a la k of
methodologies apable to provide dire tions and best pra ti es to innovation, as
well as a theoreti al systematization of the notions related to BI, often onsid-
ered more an art than a s ien e. In last years several tools and heuristi s have
1
This work has been partially funded by the European Commission through the
Proje t BIVEE: Business Innovation and Virtual Enterprise Environment (Grant
Agreement No. 285746).
been proposed as solutions to support an innovation pro ess, even though they
are mainly based on suggestions and 10-best-rules lists derived by personal
experien e of business experts or innovation guru. Su h attempts, although not
always parti ularly ee tive, share the idea that business have to master the
variables behind the innovation pro ess.
This work is a ontribution to investigate the adoption of a pragmati and
systemati approa h to support business users in the management of an innova-
tion pro ess, with the aim to drive business innovation from the art towards
the s ien e side, by also indi ating solutions that are already available for spe-
i s ienti elds, espe ially in e-S ien e. In more details, Se tion 2 dis usses
the main similarities and dieren es between a s ienti pro ess and an innova-
tion pro ess, omparing the formulation of a hypothesis to the de nition of an
innovative idea, and proposing a s ienti approa h to estimate the ee tiveness
of the latter; an open innovation perspe tive is introdu ed in the ontext of Vir-
tual Enterprises. In Se tion 3 we re ognize some fun tional and non-fun tional
requirements for an ideal BI framework, together with some basi te hnologies
and tools that an be able to support it, while in Se tion 4 we make referen e to
some spe i existing platforms supporting experimentations in e-S ien e, whose
fun tionalities an be reused or adapted to provide more advan ed support for
BI. Finally, Se tion 5 ontains some nal remarks.
2 S ienti and open approa h to Business Innovation
A rigorous and methodi al approa h distinguishes s ien e from other forms of
explanation be ause of its requirement of systemati experimentation and repro-
du ibility. The s ienti method is re ognized as the basi resear h paradigm for
understanding the validity of a hypothesis. Given the phenomenon to be studied,
this in ludes the following ma ro-a tivities, with possible iterations, overlapping
and parallelization:
Observation: the a tivity of gathering fa ts and data about the phenomenon
under study, often driven by the re ognition of an open problem or the draft
of a hypothesis.
Analysis: the a tivity of understanding, by means of manual or automati
tools, the gathered data, in order to gain insights, and possibly new knowl-
edge apable to better explain the phenomenon.
Formation of a hypothesis or a onje ture apable to explain the phenomenon,
and the formulation of a testable predi tion. Note that the pro ess often
starts with a draft of su h a onje ture.
Evaluation, whi h in ludes the planning of an experiment able to assess
whether the predi tion o urs or not.
Several elds of study and resear h, although not dire tly de nable as s i-
enti dis iplines, nd in the s ienti method an approa h to onsistently sys-
tematize and formalize their method of inquiry, in order to develop more useful,
a urate and omprehensive models and methods. In this broader sense, a s i-
enti approa h, based on methodi al management and analysis of data, ould
provide a valuable improvement for innovation pro esses in the evaluation of an
innovation idea [1℄.
Nowadays, su h a perspe tive an be ee tively put in pra ti e [2℄. In fa t,
massive volumes of data are produ ed by organizations daily: every produ t,
task, a tivity and pro ess that is planned or realized an be potentially tra ed
and logged, and this onstitutes the pre onditions on whi h an observation an
be performed. Moreover, ee tive and mature te hniques are available to analyse
data and to extra t knowledge from them.
It is to be noted that innovation pro esses, however, are in general mu h
less stru tured than s ienti experimentations, and often hara terized by non-
ontrollable or non-observable variables that ould strongly ae t the nal result.
Moreover, while s ienti investigation is aimed to dis over the stati rules un-
derlying some phenomenon in the physi al world, innovation pro esses has a
fo us on the market, whi h is dynami and subje t to ontinuous and frequent
hanges. For su h a reason, the re-exe ution of a BI pro ess, in general, may not
produ e identi al out omes. The s ienti method relies on a set of pra ti es
and formalizations whi h onstitutes a ommon ba kground for s ientists. This
in ludes, among others, methods and proto ols for evaluation and analysis, units
of measurements, standards and theories that represent a theoreti al framework
for planning and exe ution of a s ienti pro ess. Conversely, Business Innova-
tion is a less mature dis ipline, whi h still la ks enough ba kground knowledge
and theoreti al analysis to depi t a spe i methodology.
Besides su h dieren es, the two types of pro esses share some similarities.
The starting point of the investigation is represented by an idea, or a hypothe-
sis, whi h often arises from previous knowledge and the re ognition of an open
problem. Both pro esses has a dynami and risky nature, be ause the output is
not known in advan e, and ould either ontradi t or validate the idea (hypoth-
esis). Su essive iterations allows, in both ases, to make use of the urrent and
previous out omes in order to ome out with a better explained or de ned idea
(hypothesis).
A ording to su h a perspe tive, a general meta-pro ess an be devised for
innovation pro esses, in luding the following a tivities:
Observation, whi h is driven by prior knowledge or the draft of the idea,
and in ludes data trails about previous related innovation pro esses, busi-
ness pro esses, logs about internal produ ts and tasks, together with related
external data from the market, lients, and suppliers. Data about ompeti-
tors and ba kground knowledge ould be onsidered as well.
Analysis of data in order to learly re ognize open problems and opportu-
nities. In fa t, the proper de nition of an idea requires, at rst, to identify
the ause-ee t relations between the open issues to solve and the stru tural
elements of the produ t/servi e or pro ess at hand. Su h a systemati ap-
proa h ould help to point out whi h internal variables an be adjusted or
what part of the internal pro ess should be modi ed.
Formulation of an innovation idea, starting from the onsideration provided
by the previous step.
Planning of an implementation and experimentation pro ess, whi h is fol-
lowed by an evaluation phase aimed at assessing the validity of the idea in
terms of a set of indi ators and measures.
Despite its rigorous approa h, the s ienti pro ess is far from being an au-
tomatable pre-de ned pro edure to follow, be ause it strongly relies also on
imagination and reativity, espe ially for what on erns data understanding, hy-
pothesis generation and experiment planning. Similarly, an innovation pro ess
requires a deep intera tion with business users. As a matter of fa t, the role of
reativity and human evaluation in the ontext of business innovation is even
more important than in s ien e, be ause the pro ess is more strongly ae ted by
human de ision-making, and the ollaborative dimension (even within the same
ompany) is mu h more prominent.
Conventionally, the a hievement of innovation is based on the skills avail-
able within the boundaries of the ompany, and every improvement and idea is
onsidered a ompetitive advantage.Su h a perspe tive, known as losed inno-
vation, ultimately refers to ompanies as losed systems, and strongly relies on
the ontrol and ownership of intelle tual property. Re ently, also thanks to the
improvement of IT te hnologies for ommuni ation and information/knowledge
sharing, a new paradigm of open innovation is emerging [3℄. Open innovation is
based on the usage of both internal and external resour es and ideas apable to
reate opportunities for generating signi ant value. It is based on the notion
that knowledge annot be onstrained within a single ompany, team, and uni-
versity: as a matter of fa t, the availability of pre- ompetitive knowledge proved
to be apable to reate a more dynami market.Open innovation pra ti es in-
lude the exploitation of new ollaborative business models and strategies like
o-produ tion and o- reation, rowd-sour ing, peer produ tion, as well as the
usage of so ial te hnologies to support and oordinate ollaboration.
Companies ould greatly bene t from exploiting both a s ienti approa h in
the innovation pro ess and a more open attitude towards ollaboration. In fa t,
ollaborative environments, like Virtual Enterprises, where information and ideas
an be ooperatively olle ted and organized, are both a sour e and a driver for
innovation. Then, the usage of proper tools and shared methodologies within
the VE, together with a s ienti approa h to experimentation, an enhan e
the su ess rate of innovation ideas and provide a basis useful as a referen e for
future pro esses in the VE.
3 Requirements for a BI support framework
A ording to the data-driven/open perspe tive introdu ed in the previous se -
tion, for ea h of the ma ro-a tivity of an innovation pro ess we identify the main
fun tional requirements of an ideal BI framework, together with the available
te hnologies useful to provide the needed support:
Observation: the system should provide tools to support data gathering and
storage from (possibly) multiple sour es, whi h allow to have eviden es and
fa ts at disposal about the produ t or pro ess under study. Useful te hnolo-
gies in lude, besides databases and data warehouses, Customer Relationship
Management systems (CRM), Work ow Management Systems, Enterprise
Resour e Planning, market analysis.
Data analysis and de nition of an innovation idea: the framework should in-
lude support tools to analyse the olle ted observations, in order to obtain
both a summarized view of them and to extra t possible relevant hidden
relations, patterns or regularities, useful to gain new or learer knowledge
about the domain and its open issues or aws. Data Mining and Knowl-
edge Dis overy in Databases (KDD) algorithms, together with statisti al
methods are ee tive solutions for su h a purpose. Systems for ollaborative
dis ussions and knowledge sharing, then, allow new ideas and suggestions to
emerge.
Experimentation and evaluation, whi h require tools useful to support busi-
ness users in planning the innovation pro ess, in luding suggestions about
whi h steps ought to be taken in given ir umstan es, and whi h Key Per-
forman e Indi ators (KPI) should be used to gain insights about the pro ess
status. Moreover, during and after the exe ution, tools an be used to olle t
intermediate and output results, for instan e tra ing and logging systems,
CRM or surveys for information about ustomer satisfa tion and feedba k,
systems to analyse data in order to evaluate previously de ned KPIs, apable
to at h and show the impa t or the su ess of the innovation idea.
Given that in this work we refer espe ially to environments like Virtual En-
terprise's, we envisage in the following the most hallenging problems that arise
from the pe uliarities of su h an open, ollaborative and distributed s enario. On
su h a basis we ontextually de ne non-fun tional requirements for a framework
apable to support a data-driven innovation pro ess:
Integration A distin tive feature of the s ienti ommunity is the existen e
of a shared orpus of standards, norms, rules, aimed at produ ing omparable,
measurable and reliable results. Integration of knowledge from dierent s ienti
elds is feasible thanks to the usage (and the sharing) of the same ommuni ation
language, the same measurements units and ommon pra ti es and methodolo-
gies. Conversely, standards and pra ti es for BI have not been identi ed yet.
Apart ommon ba kground knowledge like logi s or statisti s, spe i business
domains may require spe i solutions. For su h a reason, a framework should
provide means to identify and des ribe resour es within the Virtual Enterprise
by referring to the same terminology, to over ome the heterogeneities among the
partners.
Complexity The la k of standard pro edures and best pra ti es ae ts also
the planning and exe ution of an innovation pro ess. The framework should al-
low to odify the dependen ies among the innovation pro ess' a tivities, provide
suggestions about whi h spe i resour e to use in a given ontext, and in the
hoi e of the best output indi ators. Complexity management also involves to
keep tra k of the status of an innovation pro ess, its variables and outputs. Com-
mon IT te hnologies in lude data management systems, optimization algorithms
and planning te hniques.
Distribution Given the data-intensive dimension of modern s ien e, espe ially
in ertain elds, a re ent trend is the onstitution of virtual laboratories, in
whi h omputation of massive datasets an be performed in a distributed man-
ner. Also an innovation pro ess ould be potentially based on several distributed
resour es, whi h are to be managed through spe i te hnologies, espe ially in
environments like Virtual Enterprise's. Besides traditional ommuni ation infras-
tru tures like internet and the Web, spe i instruments are needed whenever
the enterprise follows a more open approa h towards innovation, for instan e in
sharing of knowledge/data, of distributed tools and even of omputation, espe-
ially when innovation is highly data-driven or requires simulation.
Collaboration and oordination Sin e the ooperative planning and exe u-
tion of a omplex pro ess typi ally require several skills, both te hni al and man-
agerial, ollaboration an easily be ome a sour e of omplexity if not supported
by any kind of oordination. In last years the s ienti ommunity is showing
a ontinuously growing interest in te hnologies for data, model and work ow
sharing (e.g., [4℄), whi h onstitute the ba kbone of a more networked and ol-
laborative way to s ien e. Similarly, Virtual Enterprises ould greatly bene t of
systems to share information, data and ideas among the distributed partners,
and to support a virtual team by putting together diverse ompeten ies and
apabilities, and providing means to manage oordination and to o-operatively
perform tasks and a tivities.
4 Te hnologies for a BI framework
In this se tion more spe i solutions for a BI framework are introdu ed, taking
into a ount previously identi ed requirements to sket h up the general approa h
of su h a system, aimed at supporting business innovation pro esses.
Aspe ts of this proposal involve not only the adoption of ertain te hnolo-
gies, but also several organizational hanges. This often requires the organization
to reshape itself, and adopt a more exible attitude towards revision and im-
provement of internal pro esses and pro edures, and the on epts around whi h
the enterprise is organized. In this sense, one of the emerging solutions is the
appli ation of servi e approa h to enterprises [5℄.
Servi e orientation, in whi h single tasks and a tivities may be onsidered
as modular and (possible) distributed servi es, is apable to improve exibil-
ity and e ien y. In order to further maximize modularity and interoperability,
like in traditional SO ar hite tures, servi es an be des ribed by using the same
format, in order to provide a synta ti ally homogeneous representation of their
apabilities and fun tionalities. Su h an approa h allows to reuse some of the
methodologies and tools urrently implemented in SO frameworks, whi h an
be fruitfully exploited to respond to some requirements. One of the distin tive
aspe ts of SOA is related to the distin tion among servi e publisher (e.g., inter-
nal or external suppliers), servi e onsumer and servi e registry, where servi es
holding ertain hara teristi s or aimed at ertain fun tionalities an be retrieved
through sear hing me hanisms. Within the business domain, su h a repository
ould enable the dis overy of business servi es useful in a given stage of the in-
novation pro ess or for supporting ertain business tasks, like the evaluation of
a KPI, or the optimization of a business pro ess.
Advan ed fun tionalities an rely both on servi e and pro ess repositories
(1) to understand whi h servi es are usually applied after a given one, or (2)
to provide suggestions about whi h (typology of ) servi e is re ommended in a
ertain stage of the innovation pro ess, and (3) to dis overy the most ommon
pra ti es of usage of ertain servi es. Moreover, the des ription of su h servi es
by using semanti te hnologies to de ne a ommon terminology (at least, shared
among the members of an organization or among the partners of a VE) allows
to address integration problems.
Some general-purpose te hnologies to support ea h of the phases of an in-
novation pro esses have been introdu ed in the previous se tion. Anyway, es-
pe ially in last years, several solutions have been investigated in the s ienti
ommunity to solve similar tasks. In parti ular we refer to those s ienti elds
that are mostly on erned either with data-intensive omputation or ollabo-
rative issues, and that have at disposal advan ed tools that ould be adapted
or reused in the ontext of BI. Among them: biology and bioinformati s frame-
works, e.g. Taverna [6℄ and Kepler [7℄, support users in the design of pro esses
and work ows. Also the Data Mining/KDD domains are parti ularly a tive in
providing support for users with diverse ompeten ies in designing and exe ut-
ing a data analysis/manipulation pro ess for knowledge extra tion. Some of the
frameworks, like NeXT [8℄ or KDDVM [9℄ provide advan ed support for pro ess
semi-automati planning and algorithm/servi e mat hmaking, given that ea h
appli ative bri k of the pro ess is des ribed through some spe i language. KD-
DVM platform, moreover, in ludes a pro ess repository, useful for keeping tra k
of all the pro esses developed in the past, together with all their temporary ver-
sions. Su h a repository is used both as a referen e for next proje ts, in order
to retrieve information and details about past exe utions, and also to under-
stand whi h algorithm/servi e's sequen es performed better over ertain data.
my
For what on erns ollaborative platforms, while Experiment proje t [4℄ is
aimed at pro ess sharing within a ommunity, KDDVM provides team building
fun tionalities with fun tionalities for retrieving users with a spe i set of om-
peten ies or that were involved in a ertain past proje t. Some of su h solutions
an be reused or adapted for this purpose, in parti ular the support for pro ess
design, omposition and exe ution.
5 Con lusion
The dis ussion provided in this work is aimed at analysing at what extent a
s ienti and methodi al approa h an be adopted in the ontext of innovation
pro esses to estimate the ee tiveness of an innovative idea, in reasing the on-
trol over the pro ess and redu ing the risks of failure. Currently, the la k of
methodologies for BI represents the major hindran e against the a tual appli a-
tion of su h prin iples. To this aim, we proposed a set requirements and te h-
nologi al solutions that an onstitute the basis of a future framework, aimed
both to provide support to BI pro esses design and management, and a means to
devise and test theoreti al and pra ti al models and methods for BI. Ultimately,
a BI framework ould signi antly help in better dis riminating the best from
the worst pra ti es in business innovation pro esses, whi h an be onsidered as
the rst step towards the de nition of methodologies for BI. Although it ould
be useful to make innovation pro esses more e ient/ee tive within a single
organization, this is espe ially true in ollaborative environments, where shared
methodologies ould be identi ed starting from the information/pro esses that
are shared among the partners. As future work we would like both to deepen
the theoreti al analysis and to propose an ar hite ture of su h a BI framework.
Referen es
1. Thomke, S.: Experimentation matters: Unlo king the potential of new te hnologies
for innovation. Harvard Business S hool Press Books (2003)
2. Kusiak, A., Tang, C.Y.: Data-inspired innovation model. In: Pro . of the 36th
International Computers and Industrial Engineering Conferen e, C&IE 2006. (June
2006) 18
3. Taps ott, D., Williams, A.: Wikinomi s: How Mass Collaboration Changes Every-
thing. Portfolio Hard over (2006)
4. Roure, D.D., Goble, C., Stevens, R.: The design and realisation of the myexperiment
virtual resear h environment for so ial sharing of work ows. Future Generation
Computer Systems 25(5) (2009) 561 567
5. Poulin, M.: Ladder to SOE: How to Create Resour eful and E ient Solutions for
Market Changes within Business and Te hnology. Troubador Publishing Ltd (2009)
6. Hull, D., Wolsten roft, K., Stevens, R., Goble, C., Po o k, M., Li, P., Oinn, T.: Tav-
erna: a tool for building and running work ows of servi es. Nu lei A ids Resear h
34 (2006) 729732
7. Altintas, I., Berkley, C., Jaeger, E., Jones, M., Ludas her, B., Mo k, S.: Kepler:
an extensible system for design and exe ution of s ienti work ows. In: Pro . of
16th International Conferen e on S ienti and Statisti al Database Management,
Santorini, Gree e (June 2004) 423424
8. Bernstein, A., Dänzer, M.: The next system: Towards true dynami adaptations of
semanti web servi e ompositions. In: Pro eedings of the 4th European onferen e
on The Semanti Web: Resear h and Appli ations. ESWC '07, Berlin, Heidelberg,
Springer-Verlag (2007) 739748
9. Diamantini, C., Potena, D., Storti, E.: A Semanti -Aided Designer for Knowledge
Dis overy. In: Pro . of the IEEE International Symposium on Collaborative Te h-
nologies and Systems, Philadelphia, PA, USA, IEEE (May 23-27 2011) 8674