Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
Page 1 of 14
Shibusa: Framework to Enable Enterprise
Level Agile Adoption and Drive
Author: Swati Garg, Co-author: Zaheerabbas Shaukatali Contractor
Consultants, Agile Center Of Excellence, Wipro IT Business
Abstract
This paper elaborates a framework based on
experiential learning of coaches from Agile
Center of Excellence at Wipro. This framework
can enable large organizations to embrace Agile
in a systematic and dynamic manner. It
addresses key questions on how an organization
can make all projects get benefitted from Agile
methodologies irrespective of myriad
technologies and variable business scenarios.
Through this approach senior management
including Chief Technologies Officer
(CTO)/Chief Information Officer (CIO)/ Chief
Quality Officer (CQO) can take objective look at
progress of Agile momentum in the organization
on a periodic basis and direct their organization
to reach the pinnacle of Agility for business
enablement.
Introduction Since 2004, Agile Center of Excellence (COE)
at Wipro is actively involved in facilitating and
coaching the project teams that work on Projects
using Agile Methodologies- XP, Scrum, AUP
and so on. These projects are done for various
customers in different domains using multiple
technologies.
We, as mentors from Agile COE, consistently
observed that practitioners find it easier to adopt
when they are asked to experiment with a few
Agile practices and then gradually guided to
embrace all of Agile practices with applicable
nuances for given business scenario. All Agile
projects are not the same. The amount of Agile
practices used in different project scenarios can
vary depending upon business dynamics,
limitations imposed by certain technologies or
the culture of project team for example.
Through due course of evolution, we have
developed a framework to enable enterprise
level Agile adoption and drive at Wipro. This
framework relies on consolidation of
information based on Agile Coefficient which is
an objective measurement of respective projects
in terms of their capability to embrace Agile
practices. It helps Agile COE to stay focused
and guide the team to aim for fullest use of
Agile practices in their project scenario to gain
desired benefits. This framework is flexible to
accommodate heterogeneous projects. It
provides consolidated picture on Agile status of
all projects in the organization. Senior managers
including CTO/CIO/CQO can see how coaches
and mentors from Agile COE are contributing to
make an organization level Agile penetration
increase over a period. Senior managers
including CTO/CIO/CQO can leverage this
framework to set vision for Agile spectrum in
the organization.
Origin of Framework In order to enable Agile adoption and drive at
organization level, it is absolutely essential to
conceptualize an integrated and flexible
approach that is rooted into objective
measurement parameters to assess respective
projects. It should be easy to comprehend and
Page 2 of 14
scalable to fit heterogeneous projects. The
parameters should create consistent inference
when applied across several projects within
organization, so that senior managers including
CTO/CIO/CQO can rely on pictures produced
using the same over period of time.
Agile methodologies hinge on definite set of
practices that drive excellence of software
engineering along with adequate focus on
human aspect of team and customer
management to achieve best possible returns on
investment through an IT project. Therefore, it
is crucial to ensure that measurement parameters
must objectively evaluate the amount of Agile
practices used in respective projects. For
example: we should be able to differentiate
between a Scrum project doing once a week
builds to another Scrum project doing daily
builds. We should be able to differentiate
between an Agile project Team doing offline
code reviews to another Agile Team following
pair programming. In both of these cases the
former project has less agility as compared to
another.
Next important aspect is to ensure that different
nuances of respective Agile practices are
considered. For example a project having two
week iterations is at higher level of agility with
respect to itself when it was following 6 week
length of iterations, though both scenarios
represent an Agile project with time boxed
iterations.
In summary, Agility of a project depends on
amount of Agile practices applied and the depth
/nuance of each practice pursued for project
execution.
Agile Coefficient The Agile Coefficient is an objective measure of
project’s agility with respect to amount of Agile
practices applied and the depth of each practice
pursued. A project that leverages all Agile
practices to fullest for business enablement is
called a core Agile project. This kind of project
is represented by 100% Agile Coefficient. There
could be several projects that may have chosen
to use different and partial Agile practices, this
kind of projects may not have 100% Agile
coefficient, however, they have moved away
from conventional waterfall represented by zero
Agile coefficient. These projects with help of
Agile coaches can continue to increase the Agile
Coefficient and reach the desire level of Agility
for better business returns.
Shibusa Fundamentally we have developed this
framework based on measurement of Agile
Coefficient. We recognize striking results that it
brings to enable enterprise level Agile adoption
and drive. Nevertheless, this framework itself
can be improved further as we continue to learn.
Hence it is called Shibusa. Shibusa (noun) are
Japanese words that applies to a wide variety of
subjects and refer to several fundamental
qualities including balance of simplicity and
complexity, in a way that one constantly finds
new meanings and enriched understanding that
cause its value to grow with passing time.
Demonstration of Shibusa
Framework In this section we are moving step by steps using
sample information to demonstrate the Shibusa
framework at high level.
1. In the beginning Agile COE identifies
number of projects to pilot the framework.
For example: 100 projects out of 1000
current projects in the organization.
2. Agile COE interacts with key project
stakeholders to understand the current level
of Agile practices followed in respective
projects as identified in previous step. Based
on the understanding and evidences, Agile
Page 3 of 14
COE usages Agile Coefficient measurement
system which is explained in subsequent
section of this paper, to objectively obtain
the Agile Coefficient of respective projects.
3. Data with respect to Agile Coefficient of
respective projects is captured and
consolidated. In the beginning, a few
projects stand at zero Agile coefficient.
Some other projects reflect different Agile
Coefficients ranging from more than zero to
less than 100% respectively. For example:
Project ABC and four more projects have
25% agile coefficient. Project LMN and 7
more projects have 40 % agile coefficient.
Data emerges like the sample captured in
sorted table (Table 1) below:
Table 1
4. Consolidation of above mentioned
distribution emerges as table (Table 2)
below:
Table 2
5. Agile COE creates a consolidated picture by
plotting the distribution, while having x –
axis to represent the Agile Coefficient as
percentage, y –axis as Count of Projects.
The baseline curve is created by connecting
the distribution points starting from first
point that is created for the Project’s count
with least Agile Coefficient and continuing
to move towards 100% Agile Coefficient
cutoff as marked on y-axis. A sample
distribution plot is depicted in the figure A.
This plot represents the initial baseline of
Agile penetration in the organization with
respect to number of pilot projects selected
in step1.
6. Subsequently, in consent with project
stakeholders, Agile COE identifies areas of
Page 4 of 14
improvement with respect to Agile practices
in some or all projects as selected in step1.
Agile COE coaches the implementation and
assess the level of Agile practices once
again to plot the distribution of Agile
Coefficient. In all possible cases, the number
of projects with low Agile Coefficient
decrease and the number of the projects with
higher Agile Coefficient increase. A sample
distribution curve of improvement in Agile
Coefficient with respect to initial baseline
(Figure A) is depicted in the figure B.
7. Agile COE can overlays the figure A and B
to display the impact with respect to
increase in Agile penetration within
organization. A sample is depicted as figure
C below.
8. The improvement in Agile penetration is
quantified by plotting the distribution of for
deferential improvement in Agile
Coefficient of respective projects and
calculating the area under the curve. A
sample curve for area calculation is depicted
in figure D.
9. Agile COE can continue to repeat the step 6
to step 8 to achieve a level of Agile
penetration where no projects as identified
in step1 remain with low Agile Coefficient.
A sample of shift in curve due to
improvement with respect to figure B is
depicted as green curve in figure D.
10. Steps 1 to 9 can be repeated to include more
number of projects under influence of
Shibusa framework for Agile adoption and
drive at enterprise level.
11. The ultimate goal of Shibusa framework is
to bring all projects within the organization
represented as a dot on 100% Agile
coefficient, as depicted in figure F.
Page 5 of 14
12. Nevertheless, to be realistic CIO/CQO of an
organization can decide to operate in
appropriate spectrum of Agile Coefficient
closer to 100%. The Agile spectrum of an
organization is mainly constrained based on
landscape of its IT business.
Measurement of Agile Coefficient This section provides detailed insight on how the
Agile Coefficient is measured using aspects of
several practices that make a project objectively
move on scale of agility to meet its business
needs. When the same measurement criteria is
applied across several projects across
organization and plotted as explained in
previous section, it helps the whole organization
to gain Agile momentum.
At high level there are 33 different measurement
criteria. For sake of simplicity we have given
equal Influence Factor of 3 points for each
criterion to create a scale of 99 points or 99% of
Agile coefficient. We have set the limit at 99%
to accept the fact that 100% Agile Coefficient is
the state of perfection, which is really tough to
achieve. (Some reader may find it simpler to
make one of the criteria to have 4 points
Influence Factor and make the scale to 100%).
Each criterion has 3 or more options for
appropriate relevance in different project
scenario. Options are prearranged on increasing
scale of agile enablement ranging from zero to
five or less.
To measure Agile Coefficient of any project,
respective criterion is qualified with closest
applicable project scenario through listed
options. The Applicable Influence Factor of
Criterion (AIFC) is derived based on the
Influence Factor Multiplier (IFM) of selected
option with respect to Highest Influence Factor
Multiplier (HIFM). HIFM for each criterion
depends on number of options used for its
business scenarios. HIFM is a number less than
the available number of options (Number of
options for a criterion -1). Finally the aggregate
Agile Coefficient is achieved by summation of
Applicable Influence Factor (AIFC) of all
criteria.
Let’s take an example: A project find option C
of criterion One applicable to itself. Then the
Influence Factor Multiplier of this option (2) is
considered with respect to Highest Influence
Factor Multiplier (HIFM) as (3). Hence the
Applicable Influence Factor for Criterion One
comes as (3X2/3= 2). It finds option B of
criterion two applicable to itself. Then the
Influence Factor Multiplier of this option (1) is
considered with respect to Highest Influence
Factor Multiplier (HIFM) as (2). Hence the
Applicable Influence Factor for criterion Two
comes as (3X1/2= 1.5).
On the similar lines suitability of different
options in respective criterion is captured in the
table (Table 3) as part of this section
subsequently for further understanding of
readers who are interested in detailed approach.
Ultimately the project under example gets an
Agile Coefficient of 55.9 point or 55.9% which
is certainly better than conventional waterfall,
yet has a significant room for improvement from
an Agile Coach’s perspective. Agile Coefficient
of such projects can be improved by based on
stakeholders agreement in a given business
scenario.
Even so, the whole set of criteria and options are
open for refinements and necessary changes
based on organizational focus and needs. Agile
community can ponder collectively to give it
shape as more flexible and common
understanding framework for enterprises.
Page 6 of 14
Table 3: Detailed demonstration of measurement criteria, application options and calculation of Agile
Coefficient in an example project context
S.
N
o.
Influence
Factor
Multiplier
(IFM) of
respective
options
IFM of
selected
option
Applicable
Influence
Factor for
Criteion
(AIFC) =3*
(IFM/HIFM)
A One time development in clear phases with formal gate keeping for
each phase. Phases can include- requirements phase, design phase,
coding phase, testing phase and so on. Deployment to production
only once in the end of project.
0
B Incremental development in clear phases (with next increment and
it's phases at some lag than earlier one) with formal gate keeping
for each phase. Phases can include- requirements phase, design
phase, coding phase, testing phase and so on.
1
C Iterative and incremental development with overlapping phase yet
clear iterations of duration ranging from more than 1 month to 4
months. Output of one or more iterations ready for production
deployment cycles.
2
D Development in iterations of one month or lesser duration with no
phase demarcation yet capability of deployment to production at the
end of each iteration.
3
A One time requirements collection phase. Constant set of
requirements. Captured in a free flow document form as detailed
sections, sub sections and paragraphs . With direct access to less
than 50% project stakeholders (Mainly the requirements and design
phase stakeholders).
0
B One time requirements collection phase with tail period
overlapping to future phases. Requirements changes expected up to
50%. Captured in a tabular form of document or spreadsheet with
columns for vital attributes such as identification number,
description, priority, dependencies id, author, estimated effort/size
to implement, etc. With direct access to more than 50% project
stakeholders.
1
C Requirements collected until beginning of last iteration for
completion of project. Requirements changes expected up to 95%.
In an Agile enabled requirement tool with vital data points, and
capability to track till closer. With direct access to 100% project
stakeholders.
2
2 2
What is the project implementation strategy?
How are requirements captured?2
1 1.5
Criteria Details
Note1: Each criterion has maximum influence factor of 3 points
Note2: Each criterion has three or more options based on closest suitability
to project scenario
Note3: Highest Influence Factor Multiplier (HIFM) for respective criterion
depends on number of options. HIFM =(Number of options -1)
1
Continued…
Page 7 of 14
S.
N
o.
Influence
Factor
Multiplier
(IFM) of
respective
options
IFM of
selected
option
Applicable
Influence
Factor for
Criteion
(AIFC) =3*
(IFM/HIFM)
A No prioritization, all requirements are treated equally and signed
off to be implemented once for all.
0
B Requirements are prioritized by customer and signed off to be
implemented incrementally. Yet minimal scope for change in
priority.
1
C Requirements are prioritized by customer and team in
collaboration. Project stakeholder have common understanding for
implementation in current iteration. Yet priority of non-
implemented requirements can change any time.
2
A Based on Project Manager's experience. Completely ownership
with Project Manager.
0
B Some in-house estimation tools. Based on historical data bases and
a few project leads inputs are taken. Ownership shared by project
leads and Project Manager.
1
C Estimation sessions are live in presence of whole team. Ownership
shared by complete team and managers.
2
How comprehensively the requirements are captured from estimation point of view?
A More than 50% of project team finds it difficult to estimate due to
ambiguity and non availability of experts.
0
B Less than 50% of project team finds it difficult to estimate and
experts are available for clarifications to sort out ambiguities.
1
C 100% of project team is comfortable with estimation and there is
no ambiguity.
2
How comprehensively the requirements are captured from testing perspective?
A More than 50% of project team finds it difficult to write system
test cases/acceptance test cases due to ambiguity and non
0
B Less than 50% of project team finds it difficult to write system test
cases/acceptance test cases and experts are available for
clarifications to sort out ambiguities.
1
C 100% of project team is comfortable to write system test
cases/acceptance test cases and there is no ambiguity.
2
Criteria Details
Note1: Each criterion has maximum influence factor of 3 points
Note2: Each criterion has three or more options based on closest suitability
to project scenario
Note3: Highest Influence Factor Multiplier (HIFM) for respective criterion
depends on number of options. HIFM =(Number of options -1)
How are the requirements prioritized for implementation?
How is project estimation done?
5
6
2 3
2 3
3
4
2 3
1 1.5
Continued…
Page 8 of 14
S.
No
.
Influence
Factor
Multiplier
(IFM) of
respective
options
IFM of
selected
option
Applicable
Influence
Factor for
Criteion
(AIFC) =3*
(IFM/HIFM)
A With technical details for completeness. 100% details are
documented. Access is given to less than 50% project team. Team
does not start coding or testing activities unless design is signed off
by customer. Possibility to change is ruled out.
0
B With technical details for modularity. 100% details are
documented. Access is given to more than 50% project team. Team
does not start coding or testing activities unless design is
completed for current increment. Possibility to change exists in
next increment.
1
C With technical details for modularity. Key aspects are documented.
Access is given to 100% project team. Team do design, test and
coding asynchronously. Possibility to change exists throughout.
2
A Complex architecture and design with less than 20% opportunities
for modular changes.
0
B Object oriented, modular design up to 70% opportunities for
change.
1
C Object oriented, modular design. Not restricted for change. 2
How often re-estimation is done and what is the scope of re-estimation? On what basis?
A Less than once a quarter, as requirements are considered frozen.
Scope of re-estimation includes the complete project scope due to
large impact on work already completed.
0
B Once or twice a quarter based on change in requirements. Scope of
re-estimation includes possible impact on work already done, yet
mainly focused on remaining left over work of project based on
remaining increments of size more than 8 weeks each.
1
C Once a month or higher frequency. Minimal impact on work already
completed. Re-estimation is done for next small increment of work
to be finished in iteration.
2
To what level the Project plan is created and milestones are modified at what frequency? On what basis?
A Made at high level such as tasks are assigned in terms of minimum
1 person week and ranging up to several person weeks. Milestones
modification is almost not possible.
0
B Made at high level such as tasks are assigned in terms of person
days ranging beyond 10 person days. Milestones modifications are
possible with reasoning and negotiations, but not end date.
1
C Made up to low level clarity such as some tasks ranging beyond 40
hours. Milestones modifications are possible with stakeholders
understanding, but not end date.
2
D Made up to low level clarity such as very few or no tasks ranging
beyond 40 hours. Milestones modification is possible with
stakeholders understanding. End date can shrink or expend based on
emerging business needs and scenarios.
3
Criteria Details
Note1: Each criterion has maximum influence factor of 3 points
Note2: Each criterion has three or more options based on closest suitability
to project scenario
Note3: Highest Influence Factor Multiplier (HIFM) for respective criterion
depends on number of options. HIFM =(Number of options -1)
1.5
2 2
0
7
8
How are architecture and design created?
How object oriented/modular is architecture and design?
0
1
1 1.5
9
10
Continued…
Page 9 of 14
S.
No
.
Influence
Factor
Multiplier
(IFM) of
respective
options
IFM of
selected
option
Applicable
Influence
Factor for
Criteion
(AIFC) =3*
(IFM/HIFM)
A Available to high level stakeholders such as project manager and
customer for status updates.
0
B Published and available to next level stakeholders such as project
leads. They can contribute to status updates besides Project
Manager and Customer.
1
C Accessible and understood by all project stakeholders. All of them
have direct access to update the status of their work.
2
A More than 200 members 0
B More than 50 members yet less than 200 members 1
C More than 7 members yet less than 50 members 2
D Less than 7 members 3
A Spread across different offices in different towns, in different time
zones with difference with no overlapping work hours.
0
B Spread across different offices in different towns, in different time
zones with difference up to +/- 12 hours with some overlapping
work hours.
1
C Spread across different offices in different towns within same time
zone.
2
D Spread across different offices in same town 3
E Spread across different floors in same office building 4
F Collocated at common project floor/room 5
14
A Up to 10% team is technically strong (90% resources with less
than 1 year experience)
0
B Up to 40% team is technically strong (60% resources with less
than 1 year experience)
1
C Up to 80% team is technically strong (20% resources with less
than 1 year experience)
2
D Up to 100% team is technically strong (0% resources with less
than 1 year experience)
3
A Up to 10% team is familiar with business domain 0
B Up to 40% team is familiar with business domain 1
C Up to 80% team is familiar with business domain 2
D Up to 100% team is familiar with business domain 3
A Mostly a challenge. Rookies are expected to ramp up by
themselves.
0
B Done through specific training sessions for rookies. 1
C Done through specific training sessions and diligent on job
mentoring by senior members and peers to ramp up rookies to
comfortable levels.
2
Criteria Details
Note1: Each criterion has maximum influence factor of 3 points
Note2: Each criterion has three or more options based on closest suitability
to project scenario
Note3: Highest Influence Factor Multiplier (HIFM) for respective criterion
depends on number of options. HIFM =(Number of options -1)
4 2.4
1 1.5
2 2
2 2
1 1
How is resource skilling done?
1
11
16
How is the Project plan's availability?
How big is the project team?
How are the team members located with respected to each other?
How is team's business domain familiarity?
How is team's technical capability?
15
1.5
12
13
Continued…
Page 10 of 14
S.
No
.
Influence
Factor
Multiplier
(IFM) of
respective
options
IFM of
selected
option
Applicable
Influence
Factor for
Criteion
(AIFC) =3*
(IFM/HIFM)
A Documents, Mails 0
B Documents/Mails /Messenger 1
C Whiteboards/Wiki/Messenger/phone/v-con/web-cams /Mails 2
A Done once a week or less than that 0
B Done once a week and need basis 1
C Done formally more often than a twice a week like daily call or
sunrise sunset calls
2
A Manager is informed of priority issues on weekly basis 0
B Manager is informed of priority issues on need basis earlier than a
week
1
C Manager is proactively involved to catch and resolve the issues on
daily or real time basis
2
A Status report is sent and discuss once a week or at lesser frequency 0
B Status report is sent weekly and customer call twice a week or more
often
1
C Status report and customer call done daily once minimum or
customer is available of real time updates
2
A Documents / Mails 0
B Spreadsheets based risk resolution with specific risk attributes. 1
C visual chart on display like story boards, Kanban, Burn down charts
or others
2
A Unit testing is done after a code module is done. Formal testing
does not start until coding and unit testing (CUT) phase is not
completed.
0
B Coding and unit testing (CUT)cycles repeated at nearly a month's
frequency to complete the whole coding. Formal testing phase legs
behind by almost one month or more.
1
C Coding and unit testing (CUT) is done asynchronous with formal
testing within weekly or bi-weekly increment cycles.
2
Criteria Details
Note1: Each criterion has maximum influence factor of 3 points
Note2: Each criterion has three or more options based on closest suitability
to project scenario
Note3: Highest Influence Factor Multiplier (HIFM) for respective criterion
depends on number of options. HIFM =(Number of options -1)
1.5
1 1.5
2 3
2 3
Is there some preferred team collaboration tool?
How is the status updates made at team level?
How the status is updates done at Project Leads/Project Managers level?
How is the status update done at customer contact level?
How are results and issues tracked?21
22 How is the development cycle for coding and unit testing (CUT)?
1 1.5
1
3
19
20
17
18
2
Continued…
Page 11 of 14
S.
No
.
Influence
Factor
Multiplier
(IFM) of
respective
options
IFM of
selected
option
Applicable
Influence
Factor for
Criteion
(AIFC) =3*
(IFM/HIFM)
A Less than 1 build in 2 weeks 0
B About one build per week 1
C Builds as frequent as once a day 2
Builds as frequent as multiple times a day 3
A Done for all code on offline and manual basis. All review
comments are documented manually and submitted in repository.
0
B Done for critical portions of code on offline and manual basis. Rest
of the code is reviewed on sample basis. Static analysis tools are
used of some portion of code. Key review comments are captured
with help of tools.
1
C Done pair programming for online review for critical portions of
code. Rest of the code is online reviewed on sample basis. Static
analysis tools are used of whole code. Key review comments are
captured with help of tools.
2
Done pair programming throughout and static analysis tools are
also used.
3
A Unit test cases are written at the end of coding phase. Later on Unit
Testing in conducted and unit testing results are documented
separately.
0
B Unit test cases are written at the end of coding for each module.
Unit Testing in conducted at module level and unit testing results
are kept to signoff coding and unit testing phase.
1
C Coding and Unit testing at done at module level through weekly or
larger duration increments to be pushed to next phase in
incremental manner.
2
D Unit test cases are written before writing the code for respective
requirements. small increments of code are written to pass the Unit
test case. Done on daily basis with any demarcation of Coding and
unit testing phase. Mainly done automatically, thoroughly and daily.
3
A Previously written code's quality is improvised only on formal
demand and approval by customer.
0
B once in a while team takes a diligent timeout to improvised the
code that was written in past.
1
C Previously written code is refactored to improve maintainability at
ongoing basis.
2
Criteria Details
Note1: Each criterion has maximum influence factor of 3 points
Note2: Each criterion has three or more options based on closest suitability
to project scenario
Note3: Highest Influence Factor Multiplier (HIFM) for respective criterion
depends on number of options. HIFM =(Number of options -1)
1
3
1.5
2
2 2
25
26
How is unit testing done?
What is Code Quality/Refactoring approach?
23
24
How often the build is created?
How are Code Review conducted?
3
2
Continued…
Page 12 of 14
S.
No
.
Influence
Factor
Multiplier
(IFM) of
respective
options
IFM of
selected
option
Applicable
Influence
Factor for
Criteion
(AIFC) =3*
(IFM/HIFM)
A Formal integration and system testing execution start after formal
signed of overall Coding and Unit testing phase.
0
B Formal integration and system testing execution start after
completion of one or more increments of Coding and Unit testing
phase.
1
C Formal integration and system testing execution is done
asynchronously with design, unit testing, coding for different
requirements within scope of time boxed iteration.
2
A Clear demarcation of coding and testing phases 0
B Testing phases lag behind development phase 1
C No phase segregation between testing and coding activities 2
A Customer is involved for sign off on each phase deliverable such as
Requirements document, design document and so on until signoff
on system test result document and deployment document signoff.
Sometimes or often customer gets involve in User acceptance
testing phase to signoff.
0
B During Requirement and design phase, functional insight is
provided as 'Proof of concepts' demos. Functional increments are
deployed to testing environments during subsequent phases and
progress is demonstrated on demand. Deployment to production is
done after approval of User Acceptance Testing.
1
C Customer is involved in real time to experience the functionality
developed. At the end of iteration the functionality is demonstrated
and deployed in production.
2
A Requirements change anytime randomly during iteration for the
current iteration itself. Customer thinks this is Agile and team is
under continuous pressure to deliver.
0
B Requirements changes entirely on completion of iteration demo.
Customer shares responsibility to avoid reoccurrence in future
iterations.
1
C Requirements are affirmed and changes are proposed for yet to be
implemented functionality.
2
Criteria Details
Note1: Each criterion has maximum influence factor of 3 points
Note2: Each criterion has three or more options based on closest suitability
to project scenario
Note3: Highest Influence Factor Multiplier (HIFM) for respective criterion
depends on number of options. HIFM =(Number of options -1)
0
1.5
1 1.5
1 1.5
0
1
29
30
How is the customer involved for technical or functional progress of project?
What is impact of customers involvement to change in requirements within iteration?
27
28
How is Intergration/System testing done?
How is development and testing phase demarked?
Continued…
Page 13 of 14
S.
No
.
Influence
Factor
Multiplier
(IFM) of
respective
options
IFM of
selected
option
Applicable
Influence
Factor for
Criteion
(AIFC) =3*
(IFM/HIFM)
A Internally by stepping up project team as customer is rigid to
change in plan
0
B Extensive rounds of impact analysis and discussions are done get
partial agreement on dropping some requirements or adding some
more iteration.
1
C Due to thorough involvement at shop floor good understanding is
developed between customer and team for agreement on dropping
low priorities requirements to meet the budget or expend it to run
more iteration for completing changed requirements.
2
A Not conducted or conducted at the end of project as Project post
mortem analysis.
0
B Not conducted at definite period. Lesson learnt during some
increments are discussed and corrective actions are taken for
burning issues in subsequent increments.
1
C At the end of every iteration formal retrospective meeting is
conducted to discuss both good experiences and opportunities of
improvements, so that the team commits to repeat good practices
and find solution to opportunities of improvements.
2
A There are some Customer Complaints. Team attrition is more than
15% till date.
0
B There are no customer complaints. Team attrition is more than 15%
till date.
1
C There is customer appreciation. Team attrition is limited to 15%
till date.
2
D There is customer appreciation. Team attrition is less than 5%. 3
55.9
Criteria Details
Note1: Each criterion has maximum influence factor of 3 points
Note2: Each criterion has three or more options based on closest suitability
to project scenario
Note3: Highest Influence Factor Multiplier (HIFM) for respective criterion
depends on number of options. HIFM =(Number of options -1)
0 0
0 0
0 0
Aggregate Agile Coefficient of project = Sum of Applicable Influence Factor for all criteria =
AIFC1+AIFC2….+AIFC33
How is customer's satisfaction and team's satisfaction?33
31
32
How are requirements changes handled within and across iterations?
How is retrospective meeting conducted?
Key Pointer to remember about
Shibusa Framework Shibusa is a continuous and elastic framework to
enable Agile adoption and drive at enterprise
level. One can find newer ways to expend or
refine it.
Agile Coefficient is an objective measurement
of respective projects in terms of their capability
to embrace Agile practices.
When Agile Coefficient is measured and
improved in one or more projects of an
organization, it helps to improve the Agile
penetration of the organization.
Increase in Agile penetration of an organization
can be quantified by measuring the differential
improvement in Agile Coefficient for projects
that are brought under influence of Shibusa.
The Spectrum of Agile Coefficient for an
organization is largely influenced by landscape
of its IT business.
Page 14 of 14
Parting thoughts Currently at Wipro we have considerable
number of projects in 70% to 95% spectrum of
Agile Coefficient.. Over a period Shibusa
framework would continue to help us to focus
efforts for increasing Agile penetration at Wipro
and ensure that more (if not all) projects achieve
higher Agile Coefficient in the pursuit business
enablement. We are moving away from band of
low Agile Coefficient in agile manner.
Practically, an IT Services Company like Wipro
will take time to shrink the Agile spectrum (to
operate closer to more than 60% of Agile
Coefficient for most of projects) where as
existing core Agile companies may choose to
expand their Agile spectrum (Somewhat broader
than thin spectrum which is very close to 100%
Agile coefficient) if they venture in
heterogeneous IT business landscape. Shibusa
framework can be the guiding force both ways.
Your suggestions to improve the framework are
welcome. Please reach out to us.
References
Kalpana Sureshchandra, Jagadish Shrinivasavadhani,
Moving from Waterfall to Agile; Agile 2008
conference; ieeexplore.ieee.org/iel5/4599439/4599440/04599456.p
df. \ ISBN: 978-0-7695-3321-6
http://en.wikipedia.org/wiki/Shibui
Craig Larman, 2010 - Practices for Scaling Lean &
Agile Development: Large, Multisite, and Offshore
Product Development with Large-Scale Scrum - ISBN
0-321636406 (with Bas Vodde)
http://alistair.cockburn.us/Crystal+light+methods
Anthony Heath, Why switch a failing waterfall project
over to agile?
http://www.scrumalliance.org/articles/73-why-
switcha-failing-waterfall-project-over-to-agile
Mike Cohn; Patterns of Agile Adoption;
http://www.agilejournal.com/articles/articles/patternso
f-agile-adoption.html
John Hill, It's All in the Delivery;
http://www.scrumalliance.org/articles/18-its-all-in-
thedelivery
Matt Roadnight & Toby Barber, Using Scrum to
deliver a phase of a waterfall project : Experiences and
Warnings;
http://www.scrumalliance.org/resources/281