Every test suite I've ever inherited could tell me if the code passed. Almost none of them could tell me who'd notice if it didn't.
I have watched beautifully engineered suites collapse under one question: for whom, specifically, does this matter? Half the room usually cannot answer. They have tested everything and, in effect, no one.
Gerald Weinberg's old line was that quality is value to some person. James Bach and Michael Bolton sharpened it with two words: value to some person who matters. Not value in the abstract, but value to a specific someone who will notice, complain, or quietly walk if you get it wrong.
It is the difference between hoovering before your mother-in-law visits and hoovering because it is Tuesday. Same hoover, wildly different stakes, and only one version gets noticed.
The best teams I have worked with do something almost rude before they write a single test case: they name the person.
- Not "the user" but the finance controller reconciling month-end at 11pm on the last working day.
- They test the workflow that person actually does, not the one the ticket politely describes.
- "Nobody who matters will ever click that" counts as a legitimate reason to stop testing something, not laziness.
- Coverage percentages give way to a ruder, more useful question: who would ring us if this broke?
That last question changes everything. A team that can name the caller tests differently from a team chasing a badge on a dashboard. One is building for someone specific. The other is building for an audit that may never come.
A quality bar with no name on it is an average of everyone's opinion, which is another way of saying it's nobody's.
into your process?
A QA Review finds the gaps in 1–2 weeks. From £1,200.