Professional Business Analysts / Writing it down

The traceability nobody maintains

Useful in principle, abandoned in month four of most projects.

Linking each requirement to its source, its design, its build and its test is sound practice and it is genuinely valuable in contexts where somebody will later have to demonstrate that a specific obligation was met. In those contexts it is worth its considerable cost, and it should be resourced properly rather than assigned to whoever has capacity.

Outside those contexts it is usually abandoned quietly. The matrix is built at the start, maintained through the first change, and then diverges from reality during the period when changes come fastest. What remains is a document that appears to provide assurance and does not, which is worse than not having one, because somebody may rely on it.

The honest decision is made at the outset: either this will be maintained, with named responsibility and time allocated, or it will not be attempted. A half-maintained traceability record is a liability rather than a partial benefit.

Where the full apparatus is not justified, a much cheaper version often is: recording who asked for each significant requirement and why. That single link answers most of the questions traceability is invoked for, particularly the one that arises in month nine when nobody can remember why a rule exists.

The cheaper substitute for traceability has one requirement to be useful, which is that it records the person rather than the role. A requirement attributed to finance cannot be revisited, because finance is not a person who can be asked whether it still matters. One attributed to a named individual and a dated conversation can be resolved in five minutes, even years later and even after that person has left, because it gives whoever inherits the question somewhere to start.