Relevance decay is diagnosable, and the diagnosis is unglamorous. Almost every input is already sitting in a system the business runs.
Language
1. The vocabulary gap. How prospects describe the problem in the first two minutes of a call, against how the business describes it publicly. Recorded already, read for this purpose almost never.
2. Description consistency. Whether the site, the profiles and the third-party listings say the same thing. Contradiction does not average out, it lowers confidence in all of them.
Evidence
3. The date on the newest proof. If it is more than eighteen months old, decay is already running.
4. Whether the published work reflects current capability. A portfolio that undersells anchors the price.
5. Whether anything independent has referenced the business recently. The slowest signal and the most valuable.
Distribution
6. Where new customers say they first encountered you. Asked directly, repeatedly, and preferred over attribution modelling for this specific question.
7. Whether the business appears at all where consideration now happens. Requires looking rather than inferring.
Expectation
8. Response time, transparency, ease of starting, and risk borne by the customer, compared against the best experience customers have anywhere rather than against direct competitors, who are subject to the same drift.
Effort
9. Effort per outcome. Touches per close, time to close, and how routine discounting has become. The single number most likely to reveal decay while it is still cheap to fix.
What makes it an audit rather than an opinion
Each item is dated, repeatable, and produces the same answer for two people looking at it. Anything requiring a judgement about whether the brand feels current is not on the list, because it cannot be compared to itself next year.
Run annually it is a maintenance task. Run once, after a decline, it is an autopsy.