Requirements
Requirements traceability: linking every requirement to customer evidence
Updated
Six months after a feature ships, someone asks why it works the way it does. Usually nobody remembers, the Slack thread is gone, and the requirement just says “the user shall…”. Traceability means every requirement points back to the evidence that justified it. This article shows what to link, how to write requirements that carry their reasons, and how to keep it light.
What traceability gives you
- Faster decisions about change. When a requirement is challenged, you can see whether its evidence still holds instead of arguing from memory.
- Fewer zombie features. Requirements with no evidence behind them stand out, and are good candidates to question or drop.
- Easier audits. In regulated fields such as healthcare and finance, showing why a requirement exists is often mandatory, not optional.
- Context for engineers. A story that links to the customer quote behind it gets built with the real problem in mind.
The chain
Insight what a customer said or did ↓ Hypothesis what we believe, and for whom ↓ Experiment how we tested it, with targets set up front ↓ Finding what we learned ↓ Requirement what the product must do
Each link points backwards. A requirement links to the findings that justify it. A finding links to the experiments, hypotheses and insights it came from. You never need to maintain the chain from the top; you only ever link to what already exists.
Stakeholder and solution requirements
It helps to keep two kinds of requirement apart:
- Stakeholder requirements state a need in the stakeholder’s own terms: “ICU nurses need to prescribe a standard infusion in under five minutes.”
- Solution requirements state what the product must do to meet that need: “The prescribing screen suggests the unit’s ten most-used infusions, fully prefilled.”
The split matters when things change. If a better solution comes along, the stakeholder need stays, and only the solution requirement is replaced.
When the evidence is not research
Not every requirement comes from discovery. Some come from a regulation, a tender, a contract or a direct customer commitment. Those are just as traceable: record the origin instead of a finding, for example the regulation and clause, or the contract and the date it was signed. What matters is that the reason is written down where the requirement lives.
Write acceptance criteria that carry the reason
Requirement: Suggest the unit’s most-used infusions, fully prefilled
Why: Finding F-12, prescribing time fell from 15 to 4 minutes
in the two-unit pilot, with no rise in corrected orders
Acceptance: - The 10 most-used infusions for the user’s unit appear first
- Choosing one fills drug, dilution and rate
- The prescriber confirms before anything is orderedKeep it light
- Link, don’t copy. Point to the finding instead of pasting its text, so there is one version of the truth.
- Approve deliberately. Have a named person approve requirements, and send them back with a reason when the evidence is thin.
- Retire, don’t delete. When a requirement becomes obsolete, mark it so and keep the history. The reason it was dropped is evidence too.
- Carry the trail into delivery. When a requirement becomes a ticket, take the evidence links with it.
Traceability starts long before the requirement, with good insights and testable hypotheses. If you are setting that up, start with the continuous discovery routine.
Keep your evidence and decisions linked
Groundwork takes your team from customer insights to evidence-based requirements. Free to start, with your whole team.
Get started