The single most useful thing about relevance decay is that it arrives in a fixed order, which means it can be caught early by anyone willing to measure the earlier steps.
The order
First, perception shifts. The business is described slightly differently, considered slightly less readily, recommended slightly less often. Almost no company measures anything here.
Second, the pipeline changes shape before it changes size. Deals take longer. More stall than lose outright. The share of enquiries arriving already convinced falls, and the sales team compensates with effort, which works and hides the signal.
Third, revenue moves. By this point the perception shift is typically a year or more old and has been compounding throughout.
Why the second stage is the best warning
It is already being recorded. Time to close, stall rate, and how convinced enquiries arrive are all in the systems most businesses already run. Nobody looks at them as decay indicators, and they are the earliest signal available without new instrumentation.
What perception measurement actually means
Not a brand survey once a year. Something repeatable: how the business is described when someone is asked about the category, whether it is named at all, and who is named instead. That is now measurable directly, which it was not for most of the history of this problem.
The discipline required
Whatever is measured has to be measured the same way each time. A perception measure that changes its method produces movement that is method rather than market, and it will be read as market. If the method changes, the comparison should be withheld rather than shown.
The test that keeps this honest
Decay is only a real explanation if it predicts before the fact. If perception measures are steady and revenue falls, decay was not the cause and something else was: pricing, execution, a competitor, a contraction. A business that attributes every decline to decay has stopped diagnosing.