Está en la página 1de 6

Are waterfall and agile

project management
techniques mutually
exclusive?
by Eve Mitchell, PwC

22 MARCH 2012 | www.pmtoday.co.uk

Change is a ubiquitous feature of modern life. Organisations


across the globe are changing their working practices
and business strategies to embrace the complexity and
interconnected nature of a rapidly changing business
environment and a shifting global economy. New delivery
models often include suppliers, customers, vendors,
partnerships and even competitors. Through these changed
structures and practices organisations are becoming more
able to address the pressures of rapid change, global
competition and increasing complexity.

Projects
need to be
managed to
be successful

The traditional waterfall project management


methodology assumes that a project is finite with a definite
beginning and end; that projects need to be managed to
be successful; and that the events affecting the project are
predictable. In addition, with this traditional methodology,
once a phase is finished it is thought that it will not be
revisited. The strengths of this approach are that it provides
clear steps for development and has a structure which
allows the project to be broken into manageable stages
for more accurate planning. However, its limitations have
been experienced by many project managers - projects
rarely follow the given sequential flow, and customers
often find it difficult to state all of their requirements early
in the project often resulting in dreaded scope creep, or
worse the project being seen as a failure as it hasnt met all
the customers requirements.
In contrast, agile project management adopts the hypothesis
that we dont know everything at the start of a project, and
even if we think we know, this can be subject to many
changes. Agile projects are structured in order to speed up
the rate of learning for the team and organisation getting
customer feedback early, identifying and ironing out technical
difficulties, testing early and often. Through a constant cycle
of learning and adaptation the agile team produces value to
the customer, and learns to do so more and more effectively
at each iteration. If the principle that each project begins with
many uncertainties and that learning is good is embraced,
we can in turn look to find ways of learning more effectively,
and to incorporate this into the project as it moves towards
its end point even to a point where we reward learning as
opposed to admonishing change.

The principles of agile

The agile methods use a variety of tactics to build a project


environment that is indeed agile. There is no one set of
practices that are definitively agile there are a dozen
different methods that are included under this banner,
each including unique practices. The Agile Manifesto
www.agilemanifesto.org describes the underpinning value
system which results in the key behaviours listed below:
l Early delivery of value
l Client centric
l Different process for delivery
l Success criteria
l Quality criteria

So, are waterfall and agile methodologies mutually


exclusive?
Through working with a range of clients across a number
of different sectors I have had the opportunity to observe
organisations who would consider they are guided by waterfall
principles unwittingly also employing agile practices.
I am sure you will recognise some of these scenarios:
l Top priority projects having a designated war room in
order to co-locate team members to allow sharing
of ideas, brainstorming, improve communications and
strengthen cross-functional relationships.
l Daily project team meetings to describe what happened
the day before and what is planned today a home
grown concept of a daily SCRUM or stand up. These
daily meetings give focus to the day and bring resolution
to issues quickly and collaboratively.
l Teams that produce work that is immediately tested
by other members of the team a continual straw
man improvement system which means that the solution
is continuously being improved and proven out.
l A team stepping into a project that has failed where
the deliverable is due to the client with no time to plan
and develop fully, therefore a real understanding of
what value the client is looking for and that what
is given with further improvements ongoing a
continual beta-testing.
Obviously these dont reflect a fully developed agile
programme, but they do represent fragments of where we
see agile in every day project management, and represent
that organisations are often willing to use agile tools in one
form or other.

Case study

To illustrate this further I would like to share a case study of


an organisation that needed to re-establish itself as a leader
in a transformed financial market. It required the early
delivery of business benefit from its new channel website,
but when the project was planned using the organisations
standard waterfall approach it showed the requirements
gathering phase would still be under way when the first
go live date was required by the business, posing a
significant commercial risk. The suggestion of using some
agile techniques was posed as it would cater for the early
realisation of business benefit through incremental delivery.

Approach

In order to make the move towards a more agile approach in


implementation the Dynamic Systems Development Method
Atern (DSDM) approach was used. DSDM is a project delivery
framework which is regarded as being agile in that it uses
concepts such as iterative and incremental development,
dynamic change control, collaboration and on-time delivery,
but was originally designed to integrate with PRINCE2 thus
more palatable for waterfall organisations

www.pmtoday.co.uk | MARCH 2012 23

Governance and control

The organisations governance structure was derived from


PRINCE2. This methodology is particularly strong in the
area of project governance as it provides an overarching
structure of governance to a project by establishing a project
board which directs the manager using management by
exception, further enhanced by positioning the business
case as the driving force of the project. However, this is not
in conflict with agile as all of these strengths stay in place.
The first challenge was to convince the management team
to change the time for releasing funding for the project.
In its current practice funding was released at the end of
start-up and initiate a project. The total project funding
had been agreed by the original waterfall business case,
but in order to support delivery funding needed to be
released incrementally. The governance board agreed to
this, allowing iterative design documents at governance
gateways. This allowed the project to commence
exploration and engineering without knowing all the
detailed requirements of the project.
As an FSA regulated business it was assumed that the risk
and compliance teams would challenge the methodology,
as it was believed that these teams would be nervous
of an agile approach as they would infer it meant less
documentation and less rigour. In order to address this
meetings were set up to show concerned parties that the
same governance gates, structures and documents would
be used the organisations waterfall documentation
were very similar to agile artefacts, just named differently.
Through the use of team rooms, enterprise tools, daily
meetings, predictive tools such as burn-down charts and
the inclusion of the business in the project team, it was

24 MARCH 2012 | www.pmtoday.co.uk

demonstrated that there was more control than on other


projects.
The use of a time box a four week time period equating
to 100 days development and 10 days testing - was used
in each iteration. Within each time box the teams would
investigate, refine and consolidate their deliverable for
that iteration, allowing them to fail fast and change
direction in a controlled way at the end of each time box
as necessary. This tool combined with visible plans which
all the team were able to talk about and describe, helped
gain the support of the governance teams. Of particular
value was the concept of failing fast as the risk team had
just audited a project that had run for three years with no
deliverable in sight.

Business need and buy-in

In order to gain buy-in, examples of where projects had


failed within the organisation due to lack of business
involvement were shared with the senior team. Based
on this, and an avoidance of using agile terminology, the
senior management team agreed that close involvement
of the business was critical to the success of the project
putting the customer at the centre of the project.
To introduce agile to the wider project group, two separate
briefings were held to explain the differences between a
waterfall and a DSDM approach. The concept of project
variables and in particular the principle of flexing the
features and not time, cost or quality were explained. A
qualified DSDM trainer was also engaged so that the entire
project team could gain a common understanding of the
approach to be used, and all gained DSDM Foundation
Certification, including business personnel.

The first
challenge
was to
convince the
management
team to
change the
time for
releasing
funding for
the project

Changing attitudes and


embracing new concepts

There was considerable resistance to an agile approach


at the start of building the technical framework as it
was perceived that everything was a must have. It was
presumed by the project team that flexing the features
would not be possible if there were time issues. However,
once examined more closely it was clear that flexibility was
possible through:
l Reduction of scale (for example through redundancy of
systems)
l Not commissioning all elements at the same time
l Not gold plating systems.

The use of
modelling
dramatically
increased
the speed
of problem
solving

Timely delivery was essential to gaining market share.


The business sponsor and project manager needed some
additional coaching to help them demote musts to
shoulds or coulds for a time box to give the flexibility of
features required in the agile method. The team were also
coached on the concept of a must for the project and
not for the time box to help the team de-prioritise some
elements of work.
As part of the initial requirements gathering the organisation
used customer focus groups to understand their customers
needs and to inform design. They also brought the same
groups into user testing. The customer would be able to see
a functioning and tested development, ready for release at
the end of each increment time box. Implementation of a
new technology framework required to support the project
would enable full incremental delivery of web changes,
supported by configuration management systems.

Collaboration and
communication

Collaboration was seen as a challenge due to the project


team being spread over three countries with different
cultures, with the business users and end customers in the
UK, and development houses in India and South Africa.
Significant attention was paid to communication issues
associated with multiple locations. To address this, the
organisation agreed to the following:

Every opportunity was taken to run facilitated workshops


when all parties were present in one location. This enabled
rich communication within the project and with the
projects customers.
The project manager or facilitator chaired all of the
meetings and made sure all team members contributed.
If decisions or discussions started to get bogged down the
team were encouraged to white board or brainstorm the
problem. Use of a dedicated team room to display project
management artefacts also improved communications
both within the project and to the whole company.
The use of modelling dramatically increased the speed of
problem solving. Once the group solved a problem they
would photograph the solution and share it with the team
for comment / feedback around the globe.

Quality

Using a Prioritised Requirements List (PRL), visible across


the enterprise, the solution testers were able to inform
the business analyst when a requirement was not good
enough to test against. Testing was also introduced during
each time box with business and end user involvement with
the understanding that full integration / regression testing
could only happen within the quarterly release cycle.

Techniques

l Regularly flying project team members to one location

Iterative Development

l Having daily stand-up calls to monitor progress using


conference call facilities where all team members were
not present

Iterative development ensured that development stayed on


track and delivered to the needs of the business:

l Using enterprise-wide tools to allow project teams to


access all project artefacts

l User experience feedback informed requirements

l The establishment of a team room for visiting team


members and UK participants.

Development was categorised as functional, non-functional


or user experience. Requirements were monitored against
a baseline, usability was tested against customer focus
groups, and non-functional requirements such as speed of
response, tested prior to deployment to live and informed
further engineering work.

Furthermore, the role of the team lead was assigned


per time box, depending on which geography the core
competency for the current part of the project development
was in.

l Wireframes for web pages informed design


l Prototyping through focus groups informed design

www.pmtoday.co.uk | MARCH 2012 25

Time boxing
Time boxing, allied with collaborative planning and
estimating sessions, was key to delivering on time and
on budget. During a time box the project team regularly
recreated new requirements, which were placed on the
PRL. At the end of the time box the PRL was assessed to
see if the teams experience would move new requirements
into the upcoming time box replacing previous plans. This
decision was always made by the business, informed by
the team.
This was a change from the organisations existing way of
working. Adding or removing requirements were seen as
a change of scope requiring a change request and impact
assessment. Working with DSDM saved at least one manday per week by not performing these tasks.
Modelling
Modelling was regularly used in facilitated workshops as a
communications technique, most commonly using a white
board or flip-chart in line with the DSDM model.
Models were:
l Processes
l Data models
l Functional diagrams
l Technical diagrams
l User stories or journeys
If the result was needed as a product it was digitally
photographed and recreated in an appropriate software
package. In some cases the images were so powerful that
they were created as storyboards and used as project
communications tools. If they added understanding to a
requirement they were added to the entry in the PRL.
Prototyping
Prototyping was used from the start of exploration to help
design web functionality and determine usability, from web
wireframes through to fully developed pages with stubs
to replicate real data. Prototyping has rarely been used in
the organisation, and the increase in pace and quality of
requirements gathering which would have normally been
done through one to-one sessions between business and
analysts was a surprise to team members.
Roles
Several conversations were held with the business about
the required seniority of the Business Sponsor in the
organisation truly being able to force open closed doors.
Furthermore whether they were willing to empower other
team members to make decisions The empowerment
conversation was challenging as the corporate culture was
hierarchical and made more difficult by the ability of many
people to say no rather than yes. The turning point in
the conversation came when there was a realisation that
through the use of time boxing any poor decision would
be highlighted within the time box cycle (4 weeks) in the
time box review which they would attend.

26 MARCH 2012 | www.pmtoday.co.uk

Lessons learned

Open communication can cause challenges


By opening up all parties to all communications and
changes we created some re-work in our first engineering
time box. This is because one of the offshore development
teams interpreted every design change to web pages
during the feeder exploration time box as a programming
change and therefore work to be done.
The communications channel between the design team
and that specific offshore team was removed, only passing
information when design was signed off.
Work being done that was not on the PRL
Any requirement needed to be entered on the PRL, whether
it was functional, non-functional or user experience. If it
was not on the PRL it was not done, and more importantly
not billed.
A week long gap appeared between time boxes which
was a one-off due to poor planning. This was filled by
exploration work commissioned by a team lead without
the knowledge of the project manager, even though daily
stand ups were happening during this period.
To ensure this didnt happen again the service requests
were re-written to call out sections of the PRL as the
only source of work. In addition, the team were coached
any planning needed to finalise the upcoming time box
occurred in the final week of the current time box.
Daily stand-ups
Fundamental to the success of this project and to ongoing
collaboration was daily communication with teams in three
countries. At least once every two time-boxes the team
leads flew to the central location for a fortnight to ensure
the team maintained the closeness that had been started
at the kick-off.
Updating the burn-down chart
In the first time box the burn down was used almost as
an attendance register, demonstrating the developers
days worked rather than a demonstration of functionality
achieved. The teams were reminded of the need to flex,
which could only be achieved by knowing what was
complete and what was not.

The
empowerment
conversation
was challenging
as the corporate
culture was
hierarchical

l Enabling an agile ethos through collaboration and


communication, allowing dynamic change and
accepting that detail often emerges during the project
and cannot be predicted at the outset
l Promoting the use of facilitated workshops and
modelling
I often find the need to flex my working practices to
incorporate different aspects of what I consider best
practice to both fit with the organisation in question, but
also to deliver value to the customer. Through combining
the best features of waterfall development disciplines with
agile principles it is possible to get superior results.
Arguably, the purpose of any project is to deliver value to
stakeholders. By embracing new and different principles
and techniques and applying them in the appropriate
setting to deliver value, a project will be well on its way
to success.

In the second time box the suppliers were amazed that


the business ambassador removed some of the work to be
done from the time box, rather than cracking the whip.
This was a major turning point in the relationship between
all parties.
Outcome
This was a pilot project for the organisations transition
to DSDM as a method of choice and was thus open to
a great deal of scrutiny. Using some DSDM techniques
demonstrated a level of control not seen in the organisation
before, which tipped the balance of opinion in favour of
using these methods. The delivery of business benefit early
in the project lifecycle meant that the companys position
in the market was strengthened, meeting a significant part
of the business case.

Conclusions

So, back to the original question, are waterfall and


agile project management techniques mutually
exclusive? Based on the above case study, I think they
are not. Integrating DSDM into a PRINCE2 environment is
relatively straightforward in that both of the approaches
are constructed in a similar way. They both have similar
organisational structures and thus when combining the
two there is no need to duplicate or compromise, more a
need to take on additional concepts and components from
agile and integrate them into the structure of PRINCE2
Notable enhancements include:
l Introducing the mechanisms for scope tolerance using
a PRL and techniques such as time-boxing
l Explicit support of iterative and incremental product
development

28 MARCH 2012 | www.pmtoday.co.uk

Eve Mitchell