Tag Archives: requirements

Applying Agile principles to requirement analysis

23 May

Background

The Agile methodology originated within the software development industry. Since its inception in 2001 – Agile has expanded beyond an initial developer-centric community – to being embraced by multi-discipline teams working across numerous industries.

The antecedent of Agile within IT was the Waterfall methodology. The Waterfall framework consisted of a series of sequential, discrete phases – with each phase conveniently mapped to a role/responsibly:

Analysis Phase             -> Requirement Analysis (Business Analysts/Product Owners)

Design Phase                -> UX (Designers/Usability Experts)

Development Phase  -> Software Development (Developers)

Testing Phase               -> QA (Manual Testers and Developers in Test)

Delivery Phase             -> Release Management (Project Managers)

Due to the increasing popularity of Agile – requirement analysis has been encouraged to transition from being a stand-alone phase owned by BAs/POs – to become a project facet that can incorporate Agile principles.

In what ways can requirement analysis adopt Agile principles?

Collaborative requirement analysis

Prior to Agile – the practice of the development team being presented with an upfront, non-negotiable, detailed requirements document (BRD/functional specification etc) was common. With the advent of Agile – requirement analysis should no longer be restricted to the interaction between BAs/POs and the business – instead we should embrace collaborative requirement analysis:

A popular collaborative requirement technique is the “3 Amigos”.  This process involves the developer, BA and QA discussing the requirement specification in a workshop. Each Amigo will offer a unique perspective – through discussions the Amigos will identify edge cases, undefined requirements, opportunities and potential reuse. The 3 Amigos technique can also reduce the risk of incomplete features being pushed into development by the product team – requirement specifications must be pulled into development when they have been reviewed and accepted by the 3 Amigos.

Collaborative requirement analysis facilitates a project-wide sense of ownership – and also communicates a common understanding of what features need to be built. Collaborative requirement analysis produces more robust specifications – and reduces the role-based silos that can exist on projects.

Detail as an emergent property

Agile artefacts such as technical spikes and development iterations mean that high-level requirements can be considered sufficient at project initiation. Low fidelity requirement assets (e.g. user stories /”back of the napkin” designs) should be employed on Agile projects:

Just-in-time requirements analysis (JITRA) has a concept that requirements should only be specified at the level of detail required for upcoming development. JITRA states that the further in advance of development requirements are defined – the more probable that requirements will become out of date, leading to rework and wasted effort.

Detail should emerge when it is required – which is typically towards the middle/end of the project lifecycle. Initial requirement analysis should be focussed on business justification and solution scope.

Embrace change

Specifications will evolve throughout the project lifecycle; all team members must acknowledge the benefit of responding to change. Adapting to changes in circumstances/urgency/understanding is crucial – requirement analysis should be considered an iterative rather than exhaustive process:

In terms of systems theory – project teams should be viewed as open systems. As the system will tend towards a steady state – change should be encouraged and communicated at an organisational level. Regular priority sessions, stakeholder workshops and competitor reviews should be used to mitigate resistance to change.

Incorporating feedback is crucial to the success of a project. Requirements are not unchangeable statements – they only reflect the current and expected situation, both of which are liable to change.

Necessary documentation

The adoption of Agile principles does not mean that requirements should not be documented. Requirement documentation is vital for developers, QA and the business stakeholders:

The principle of living documentation should be embraced. This means that all documentation needs to be accessible and up-to date. Business users, developers and QA should be able to request requirement changes. Documentation is most valuable when it is understandable by all team members, available and responsive to change.

Lightweight documentation such as feature files and high level process maps summarise the output of the requirement analysis process. The Agile methodology encourages appropriate documentation – superfluous detail is wasted effort; Agile does not negate documentation.

Continuous process improvement

Requirement processes should not be viewed as immovable obstacles. Instead these processes should evolve and adapt to meet the needs of the project. Where a process or artefact ceases to produce the expected value –it should be reviewed and changed by a self-organising team:

Retrospectives are a popular technique for identifying improvement opportunities. Team members meet to discuss what the team needs start doing, stop doing, and continue doing. Regular (every 2/3 weeks) and actionable retrospectives provide an open forum for continuous process improvement.

Requirement analysis processes (to-be-analysis, process mapping, stakeholder workshops etc) can always be improved. A technique that is effective for one team – may not be effective for another – or at least may require several modifications.

Continuous delivery

The Agile methodology promotes product iterations and regular releases. In order to align with this ethos, requirement analysis must produce a constant output – a steady flow of requirements will avoid the “big bang” requirement delivery that characterised the Waterfall methodology:

Minimum Viable Product (MVP) provides the scope of requirement analysis. The MVP will be delivered in multiple iterations – requirement analysis must be constantly baselined against the MVP and ensure that there is a sufficient specification available for each delivery.

Shorter delivery timescales encourages more frequent requirement analysis output. Specifications should be aligned to the MVP – features need to be deliverable and contribute to the MVP vision.

Conclusion

Iterative, collaborative Agile development has replaced the sequential Waterfall development methodology. Prior to Agile – the product team could hand over a list of detailed requirements – which would then be used by developers to build the product. In order to align requirement analysis with Agile development practices – the following principles need to be applied: requirement collaboration, iterative specifications, embracing change, necessary documentation, continuous improvement and continuous delivery. By adopting these principles requirement analysis will transition into the Agile world, produce better specifications and ultimately lead to greater quality products.

Advertisements

How 2 types of BA can transition from Waterfall to Agile

25 Apr

Introduction

Most BAs have experience of traditional Waterfall development. Within the Waterfall framework there are the following sequential phases:

-> Analysis
-> Design
-> Development
-> Testing
-> Delivery

Waterfall projects kick-off with an “Analysis” phase. This is designed to assess the problem and scope out a solution (as-is and to-be analysis). The project will then progress through the remaining discrete phases – until the product is delivered.

In the same way that the “Design” phase had designers, the “Testing” phase had testers, the “Analysis” phases had analysts (BAs).

After working on a number of Waterfall projects – I noticed that 2 types of BA evolved.

TYPE A (“Specialists”/”Purists”)

• These individuals were lucky enough to work in a mature BA practice. The analyst role was well established – BA deliverables were often standardised (e.g. BRD templates/peer review)
• They focused primarily on the “Analysis” phase of a project – in the same way that designers focused on the “Design” phase
• When the “Analysis” phase was complete – they moved onto another project. It wasn’t uncommon for a BA to work on multiple projects/workstreams simultaneously

TYPE B (“Generalists”)

• These individuals almost had to define their own roles. They worked in organisations where the BA role/responsibilities weren’t fully understood
• BAs were treated as generalists. They were often asked to “do a bit of everything”
• Their deliverables stretched across the project lifecycle and included: requirement specs (“Analysis” phase), wireframes/low fidelity mock-ups (“Design” phase), clarifying queries from the dev team (“Development” phase), testing the product (“Testing” phase) and creating user guides/training end-users (“Delivery” phase)
• These individuals worked on individual projects – and were often influential

How can these 2 types transition into Agile?

TYPE A (”Specialists”/“Purists”)
• These individuals can continue to focus on specifying new features (i.e. requirements/acceptance criteria). In order to maintain a purist approach they will need to work 1 Sprint ahead of the developers. Most of the work they produce will go into the following development Sprint
• Career progression -> Senior BA role

TYPE B (“Generalists”)
• These individuals can continue to focus on a diverse range of deliverables: they can produce requirement specifications, provide design feedback, facilitate product planning, support development and perform ad-hoc testing. In order to provide the developers with a constant backlog of work – they will need to spend 50% of their time on the current Sprint and 50% of their time on the upcoming Sprint.
• Career progression -> Product Owner (possibly also Scrum Master role)

Thoughts/feedback?

What is your opinion? Are there TYPE As and TYPE Bs? Where do they fit in an Agile world?

Image

Brian the Business Analyst – part 2

24 Apr

brian_the_ba_pt2

Just-in-time Requirements Analysis (JITRA)

20 Apr

Introduction

Just-in-time requirements analysis (JITRA) is a BA approach based on lean/agile/kanban practices.

The 2 principles underlying JITRA are that requirements:

  • Should ONLY be identified when they are needed; and
  • Should ONLY be defined at the level of detail required

The first principle aims to optimize the timing of requirements; requirements should be delivered based on need rather than convenience (e.g.the  development team needs this next – versus – this feature can be specified easily).

The second principle aims to optimize the cost of requirements; superfluous detail is a waste of effort – and ultimately money.

Time and costing are often finite resources on a project. JITRA aims to optimize both from a requirements perspective.

Implementation

The most widely accepted framework for implementing JITRA principles breaks analysis down into 4 major activities:

  • Initial Analysis
  • Feature Set Analysis
  • Story Analysis
  • After-action Analysis

For a detailed description of each stage – please refer to: http://cf.agilealliance.org/articles/system/article/file/1007/file.pdf

Challenges

Teams that implement JITRA often face perceived challenges from a range of stakeholders:

  • BA: “Without a buffer – requirements might not be ready on time. I don’t know how long it will take to analyse the requirements until I begin!”
  • PM: “Without detailed analysis at the start of the project – how can I estimate the delivery date!”
  • Developers: “Isn’t this just product leaving it until the last minute – and giving developers incomplete requirements!”

Although these perceived challenges could stop a team from experimenting with JITRA principles – there are strong advantages to the approach.

Advantages

  • Agile: JITRA reiterates that the Business Analyst should work on features that can go straight into the backlog. This should provide the development team with a continuous flow of requirements and avoid a “BA bottleneck”. Additionally – as requirements don’t need to be fully specified upfront – JITRA enables requirement details to emerge during iterations.
  • BA perspective: the further in advance of development that requirements are defined – the more likely they are to become out of date. This in itself will lead to rework and ultimately require more analysis effort. With JITRA all three of these issues should be addressed.
  • Quality: the closer to delivery requirements can be left – the more information a BA has on which to build. This will lead to more valuable requirements.
  • Developers: Product requirements typically exist at a high level – the BA provides the detail. One problem with this approach is that if a BA provides details far in advance – one or two specific (and probably minor/low value) detailed requirements could cause considerable development challenges. If the BA provides initial high level requirements – then the development team can present back a set of options for  detailed requirement – the development team can also quantify these options in terms of effort/risk/ technical elegance.
  • PM: This approach requires a smaller upfront investment from the BAs. It also reduces waste from the requirements stage  – as there is less BA rework and less redundant effort.