-The value stream mapping is one of the lean practices, that helps to visualize the whole process and identifies any bottlenecks affecting the flow. Proper management of the value stream can significantly contribute towards waste elimination by categorizing process activities to be either value adding or non value-adding. Lean development focuses on the value through the elimination of waste. Adding value through embracing change and customer satisfaction are also the benefits of Scrum. This study reports our experience regarding the implementation of VSM with Scrum. We followed the action research method, with an objective to see if VSM can contribute to the identification and reduction of wastes in a Scrum-based project. We identified a noticeable amount of waste even with strict compliance to the Scrum practices. On the basis of identified waste, their root causes, and possible mitigation strategy we have proposed a future state map, that could help improve the productivity of the process. The results of our study are encouraging, and we suggest that adoption of VSM with Scrum could add more value to the Scrum-based projects.
The focus of software development organizations is to
develop high-quality software at a lower cost, faster pace,
and maximum customer value. To achieve this, organizations
have shifted their approach from classical development to
iterative releases. In early 90’s people started raising their
voices to curtain lengthy processes and documentation in
software development. Agile and lean are the methodologies
which supported this concept. Both Agile and lean have their
own well-defined set of principles. Some of the principles are
shared among the two, at the same time each has its unique
principles and practices [
Scrum is a popular agile development method. It is well
suited where requirements are random and complex [
The overall goal of lean development is to achieve a
continuous and smooth flow of production with maximum
flexibility and minimum waste in the process. All activities
and work products that do not contribute to the customer value
are considered waste. Identifying and removing waste helps
to focus more on the value creating activities [
One of the important lean practices is value stream
mapping(VSM). It is the process of directly observing the flow
of items, and summarizing them visually [
This study briefly describes the main concepts associated with VSM and presents an analysis of VSM implementation in a Scrum-based project. The focus of the study is to see if the VSM can be beneficial when used in Scrum. In our project, by identifying the value-added and non value-added time in the current state of the process, we were able to identify and prioritize the various type of wastes. Furthermore, we have proposed the mitigation strategies to reduce the identified waste and thus suggested a future state of the process. Our experiment shows that the use of VSM with Scrum adds more value when comparing to the effort utilized on it.
The rest of this paper is organized as: Section II describes the background knowledge and concepts regarding value stream mapping. Section III elaborates on the method used for the study, it also enlists the research questions investigated in the report. VSM process execution is presented in Section IV. Section V presents the findings of the research. Finally, conclusions are provided in Section VI.
This section presents the background knowledge and key concepts regarding the value stream mapping. The subsequent sections describe the following concepts: A) The value stream mapping, and its purpose, and B) How to implement VSM.
Womack et al., [
Value stream mapping, if implemented properly, can have
significant contributions towards waste identification and
increased process efficiency [
What is Value? Value is the end-goal and therefore the
establishment of value parameters at the start of a project
is the key to achieve improved productivity and customer
satisfaction. Emmitt et al., [
The perception of value is individual and personal, and is therefore subjective. Agreement of an objective best value for a group will differ from the individuals’ perception of value.
Values will change over time.
These characteristics highlight the complexity of value and thus emphasize on having a clear understanding and establishment of value before doing anything else.
The Value Stream Rother and Shook [
The Flow and Flow Items: To look for a bottleneck in a
production system, it is important to understand what flows
through that system i.e. the flow items. All work that flows
through a software value stream is characterized by one, and
only one, of the following flow items[
Features and Defects: If we consider feature additions
and defect fixes as the flow items. We can characterize
work across all the people and teams in a value stream
as applying to one of these units. Having visibility into
every process, we could identify exactly how many people
were involved in creating, deploying, and supporting a
particular feature. The same goes for a defect fix. [
Debt Reduction: The fourth type of work that can be
observed is debt reduction. If automation is not done
to reduce infrastructure debt, it could impede the future
ability to deliver features. This work tends to be pulled
by software architects.[
How to Measure Flow? A number of metrics can be found in literature that can be used as measures of software delivery flow. They include lines of code, function points, work items, story points, deployments, and releases.
Each of them captures a notion of value flow from a different
perspective, and has its limitations when used to depict end to
end flow. [
Metrics for VSM: Following are the metrics that are
associated Value stream mapping and can be used as a reflection
on the current state of a process[
Takt time: Takt time is the rate at which a company must produce a product to satisfy customer demand.
Cycle time: The time that elapses from the beginning of a process or operation until its done.
Total cycle time: The total of all cycle times for each individual operation or cell in a value stream. Ideally, if there is no waste then the Total cycle time equals total value-added time.
Queue time: The time that a work unit will wait for a downstream operation to be ready to work on it.
Lead Time: Number of minutes, hours, or days that must
be allowed for the completion of an operation or process,
or must elapse before a desired action takes place.[
Value stream management is a process for planning and
linking lean initiatives through systematic data capture and
analysis. Value stream management consists of eight steps
[
The first step to improvement is willingness to change. Successful VSM implementation requires a commitment from management to maximize the visibility of operations and accept any changes as suggested by this process visualization activity. 2) Choose the value stream:
A flow item must be selected as a target for improvement. This will define the scope of the VSM process. 3) Learn about lean:
Having the knowledge of key lean concepts, such as maintaining a consistent work-flow, the lean waste and its types, and the importance of continuous improvement (kaizen) etc, will help to categorize the activities as value adding or non value adding, hence identifying the waste. 4) Map the current state:
The current state is determined by gathering information on current practices, and a current state map is drawn by visually presenting this information.
On the basis of the current state map, the values for the lean metrics (i.e the lead time, the cycle time, the queue time) are derived to reflect the current efficiency of the process. 6) Map the future state:
Highlighting the non value adding activities helps to identify and prioritize wastes. A future state map is presented after addressing the root causes of waste with an intention to improve the process.
The Development of current and future state is an overlapping effort. Future state improvement ideas will come while working on current state, and working on future state may highlight important current state ideas that have been overlooked. 7) Create Kaizen plans:
The next step is to devise plans for achieving the future state by incorporating the changes as suggested in the future state map. 8) Implement Kaizen plans:
The final step is the implementation of Kaizen plans that describe how to achieve future state and how to cope with this transformation.
We adopted action research as main research method for this
study. Action research is a participatory approach concerned
with developing practical knowledge [
The research activity is based on research questions presented in Table I.
To extract knowledge about the VSM, we carried out a literature review and studied literature presented in various research articles. After having a clear understanding of VSM, we utilized the acquired knowledge to observe the outcomes
Identification of flow items
Data collection
Current state mapping Identification and prioritization of waste
Future state mapping of value stream management in a practical scenario. The steps taken for this study are illustrated in Figure 1.
The VSM activity was performed on a Scrum based process. The first author of the report was the member of the team executing the process, and served as a Scrum Master. There were two iterations in the process, but the perception of value changed for the customer after the demo of first sprint and the project scope was redefined. The tasks delivered at the end of the fist sprint were rendered as waste as the customer was not going to utilize the outcomes of that work. The sprint under observation consisted of 12 working days, and the value stream was mapped for the same period. A working day was considered to be of four hours as the team members were allocated fifty percent of their capacity to this project.
For the implementation of VSM, we followed the steps
defined by Tapping et al. [
The first step is to determine the flow items for the value
stream and use some measures for them. Kersten [
An online Scrum management tool was used as a virtual Scrum board, for visibility and availability purposes. This tool was also used for time based data collection. All the team members reported the actual time using the tool every time they worked on a task. This input was further validated in the daily stand-up meetings. The data on actual time spent on each task was also maintained in an excel file, which depicts the day to day progress (See Table II). The tasks are colorcoded which makes it easy to visualize the progress of each task across different operations in the process. Time spent on each task was recorded, and if there were multiple tasks in an operation on a particular day, time spent on task switching was also recorded. There are columns to show the time for which tasks waited until they were retrieved by the next operation.
Presenting a detailed picture of sprint execution,the excel file can be used to conduct an analysis and identify the value added Time (VAT), as well as the non value added time (NVAT) per task, for each working day. The Data Collected during this phase is summarized in Table III.
This activity is completed in two steps, firstly on the basis of analyzed VAT and NVAT, we calculated the required metrics for VSM and then we generated the current state map.
a) VSM Metrics: 1) Cycle Time: The cycle time is a measure of the time required to complete one cycle of an operation or to complete a function. It consists of both value added and non-value added time within an operation. Value added time (VAT) is referred to the time spent on the activities that add some value to the customer. Whereas, non-value added time (NVAT) is the one that is spent on activities that could be essential for the process but adds no value to the customer. The cycle time is calculated for each task and the cycle time for the individual operations is presented as a sum of cycle time for each task. For example,referring to Table III, the analysis was performed by the product owner in one working day (i.e. 4 Man-Hrs). The cycle time calculated to be 4 hours, of which 1.48-hour is the NVAT and 2.5-hour is the VAT. In Table III VAT for PF-34 and PF-35 is zero because these tasks were unfinished from the customer perspective as these tasks did not progress to the subsequent operations (i.e., Development to QA). Cycle times for other operations is Design 12 hours, Development 28.48 hours, Code review 9 hours, and Quality Assurance 14.91 hours. 2) Average Cycle Time: As the cycle time for each task varied depending upon the complexity of the task, so an average cycle time of an operation is calculated by dividing the cycle time of the operation by the number of tasks. For instance, in the development operation the cycle time was 28.47 and the total number of tasks in the operation were 10. So, the average cycle time is calculated to be 2.8. Average cycle time for each operation is presented in Table III. 3) Total Cycle Time:
The Total cycle time for the system is calculated by adding the cycle time of all operations (analysis to quality assurance). Total average cycle time calculated for our system is 68.37 hours. 4) Queue Time: Queue time is the time for which a task waits between two successive operations. There were two queues in our system, one between development and code review and the other was between code Review and Quality Assurance. The task with the highest queue time is PF-28, that waited for code review for 13.67 hours, because of its interdependence with two tasks (i.e., PF29 and PF-30). The Queue time is calculated for each task, and the total queue time between development and code review is 20.07 hours, and the total queue time between code review and quality assurance is 4 hours.
The queue time for the whole process is 24.07 hours. 5) Lead time: Lead time is the measure of time required to perform all the activities from the initiation of a task until it is completed. The lead time includes both the queue time and the cycle time. If there is no waiting time between the operations , then the lead time equals the total cycle time for the process.
In our project the lead time is calculated to be 92.43, it is the sum of the total cycle time and the total queue time. 6) Takt Time: The Takt time is used to synchronize the rate of production with the rate of demand. It establishes a cadence for the work flow, and contributes towards efficiency through continuous adjustments. It is computed by dividing the available time by the number of tasks. There were 12 working days with four hours for each day (i.e., 48 work hours), and tasks to be accomplished were 12. The Takt time is calculated to be 4 hours (total work hours / total number of tasks). That indicates that each task needs to be worked by the team for four hours. b) The Current State Map: A Current State Map was generated based on the data presented in Table III, and is shown in Figure 2. We considered the customer as the start point for our CSM because the first step in value stream management is to establish a clear perception of the value as defined by the customer. The process of implementing the current state map was initiated as soon as an agreement was made on user stories. Each cell of the current state map represents an operation. The number of people carrying out the operation is also mentioned within each cell. An activity is considered to be value adding if it contributes to customer value, and all non value adding activities are considered a waste. The value added time (VAT) for an operation is represented by the trough of the time-line and the crust shows the waste represented as a sum of NVAT and queue time. The flow of information between cells is in electronic form.The takt time , lead time and process cycle efficiency are also presented in the CSM. The purpose of VSM is to visualize the complete value stream, and it should represent the complete work flow, right from the inception of demand to the delivery of the product to the customer. So customer is considered to be the endpoint for the VSM discussed in this study.
There are seven types of waste in Lean, as identified by
Taiichi Ohno – the founder of the Toyota production process
[
The future state map suggests improvements in the process. The goal is to prepare a possible course of action to address the root causes of the waste found in the current state of the process. In the light of solutions presented in Table IV, we reviewed the current state and suggested an ideal state of sprint execution with significant reduction in lead time, queue time and NVAT. For instance, to eliminate Queue time we can merge two operations, i.e., Code and Code Review by using the concept of pair programming. This will help to rectify errors during the development operation, and the delays between these two operations can be removed. It will also help to reduce the defect rate. Similarly, our next priority was to address the issue of task switching. For this, we can define the limit for work in progress(WIP) within an operation at any given time. Limiting work in progress can regularize the flow and will result in less multitasking and a notable reduction in NVAT within the operations. We propose to set this limit to a maximum of three tasks in development and quality assurance. It will create a pull by allowing the retrieval of a task when a successive operation is ready to work on it. Subsequently, the queue time between operations can be reduced to zero.
Finally, there will be no partially done work as there will be enough time to complete all the tasks. The suggested changes are presented in the Table V. It can be seen that the lead time is equal to the total cycle time as there are no delays between the operations. The proposed future state map is provided in Figure 3. The map includes calculations for takt time, lead time, and process cycle efficiency. The customer is the start, and the end point of the future state map as well.The cells symbolize the operations within our process, along with the details of cycle time and the number of workers for each operation. The Kaizen burst symbol represents suggested improvements. We used it between design and development, and then development and QA, to represent the improvements we have proposed in the form of WIP. Also, the flow between these operations is represented by a pull symbol instead of a push.
The first objective of this study was to investigate the capability of VSM regarding the identification of current state (waste) in a Scrum-based project. The current state map presents the current set of activities been performed during a process, and is identified on the basis of detailed study of the actual process. We applied the action research method and carried out a project with a team of six people practicing Scrum as their development methodology. Being aware of the overall process flow, the Scrum Master (first author of the study) was responsible for value stream management. Required data for VSM was extracted from an online Scrum board with a little effort, and its validation was the part of every day stand-up. The detailed analysis of activities in terms of time, and value to the customer, helped to categorize them as value adding or non value adding. Also some bottlenecks were highlighted that affected the overall productivity and increased the lead time for the process . The data of the current state is presented in the Table III, and visual representation of the current state is presented in Figure 2.
The current state map helped us to recognize the waste
within and between the operations. We categorized the waste
by using the list of waste provided by Cawley et al., [
The second objective of the study was to determine the effectiveness of the VSM, regarding waste reduction in the future state of a Scrum-based project. VSM not only helped us to identify the bottlenecks in the process, but also highlighted the root causes of the waste and allowed us to prepare a mitigation strategy. Table IV presents the prioritization of identified waste along with the root causes and mitigation strategy for waste reduction in each category. By utilizing the proposed mitigation strategy, we have created a future state map for the process. A set of data was created for an ideal state of sprint execution, see Table V, and this data was used to propose the future state. The visual representation of future state is shown in Figure 3. The future state map, shows a significant reduction in the lead time (i.e.92.43 hours to 62.58 hours), while also completing the partially done tasks in the current state of the process. This implies more productivity and thus a more satisfied customer, which is a primary goal of agile software development.
The waste is also reduced (i.e., 50.44 hours to 10.09 hours) by utilizing the team to maximize the time spent on value adding activities and eliminating the waiting time between process. The process cycle efficiency has improved visibly from 45% to 84% which implies a better utilization of available resources. The results we obtained by the implementation of VSM in our project are encouraging. Considering the difference, while comparing wastes and productivity in the current and the future state of the process, it can be stated that VSM can serve as an effective method to improve the efficiency of a Scrum team to an even higher level of performance. Value stream management allowed the team to realize impediments in the continuous flow of tasks, and to discover new ways of becoming more effective.
This paper presents a research activity employing the value stream management to determine its effectiveness as a waste reduction mechanism in Scrum. It was inferred that VSM can be effectively used to reflect the overall flow, to identify the bottlenecks and to distinguish between value adding and non value adding activities.
Significant amount of waste was identified, along with its root causes. Based on this knowledge, mitigation strategies were defined to reduce or eliminate waste so as to increase efficiency, and to deliver tasks with better quality and less lead time. Addressing the wastes can contribute towards sustainable and more predictable development cycles and can enhance the Scrum experience. It was also observed that implementation of VSM in a Scrum based project was a fruitful activity as it requires very little effort when compared to its contribution to overall process improvement.
Based on current state of the process, we have proposed a future state map. Our next focus will be to identify the steps needed to develop the kaizen plans for the implementation of the future state of our process. And then to measure the success as achieved by the implementation of these plans in a practical scenario.
The authors would like to thank all the participants of the project. This work has been supported by EASE, the Industrial Excellence Centre for Embedded Applications Software Engineering (reference number 2015-03235).