Background

'Sashimi slicing' refers to the break-up of requirements into thin vertical 'slices' that go end-to-end through the software implementation stack and produce something visible and of tangible value to stakeholders. A common mistake is to indulging in 'goldplating' of individual components that are not valuable to the customer until combined with other components. 'Goldplating' refers to over-investing in software features and may be damaging to Return on Investment.

Experience

The project team was building the system screen-by-screen and pursuing extensive finessing of each screen before moving on to the next when the system could not be meaningfully used until the core functionality of many screens was built and combined to form a complete business workflow. This screen-by-screen slicing approach proved wasteful as significant time was spent on infrequently used screen elements and cause a long customer feedback loop that hid important workflow-related issues until late in the development cycle.

Smells

• The release plan calls for a series of releases late in the project or a single "big bang" release.
• The customer deferring feedback until future iterations.
• The customer being unable to acceptance test until late in the release cycle.

Causes

• High level planning does not focus on establishing a schedule of incremental releases and identifying minimal releasable feature sets.
• Planning is heavily influenced by the way work was broken down on previous non-agile projects.
• Planning is heavily influenced by a certain specialty e.g. user experience design rather than the customer.
• There is little discussion about business value and how to realise it quickly.

Consequences

• A large amount of time and money is wasted on features or finessing of features that deliver little or no business value.
• The customer loses confidence in the project because they cannot see a meaningful product until late in the development cycle.
• The team builds a poor solution and this is not recognised until late in the development cycle when many pieces are integrated.

Guidance

Agile emphasises rapid delivery and maximization of business value. Therefore, divide work into thin slices along lines that maximise the rapid delivery of business value. Also use business value (benefit vs. cost) to guide investments in user interface build-out and finessing.

If you have had a similar experience or have a different perspective on this, please post a comment below.