Picture a Series B company that sells inventory forecasting software to mid-market retailers. This one is a composite of patterns we see again and again, not a named client. It has around $60 million in revenue, a product the customers respect, and a CMO who can recite win rates by segment without opening a spreadsheet.
Two years ago, retail operators described what they were shopping for as “inventory forecasting.” That was the company’s phrase, and it was the phrase on every analyst list. Then a smaller competitor started talking about “margin protection” instead. The math underneath was nearly the same. The promise on top was different, and it was written in the language a retail CFO actually uses.
Within a few quarters, that phrase was turning up in RFPs. When an operations VP asked an AI assistant how to protect margin on seasonal inventory, the assistant named the competitor and quoted the competitor’s framework. Our composite company was not in the answer, and the company had no idea.
The October planning offsite went well. The plan said, in effect, do more of what has been working. Thirteen months later the win rate on competitive deals began to slide, and the board wanted to know what happened to the product.
Nothing happened to the product.
Why the numbers don’t warn you
Revenue is a lagging indicator of relevance, and the lag is longer than most leadership teams assume. A deal that closes this quarter came from a shortlist assembled six to nine months ago, and that shortlist reflected what the buyer believed about the category well before that. So the dashboard you review every month is reporting on decisions the market made a year ago. When the decay finally reaches the revenue line, the cause is typically 12 to 18 months old, and reversing it means changing what buyers believe, which takes longer than it took them to change their minds in the first place.
You already understand this pattern, because you pay for it in two other places. You run a financial audit every year because errors in the books compound quietly until they are expensive. You run a security audit every year because the breach you haven’t found is the costly one. Relevance has the same shape, a slow and silent loss that surfaces late, and almost no company has an equivalent ritual for it.
What makes it dangerous is the thing we call category gravity, which is the reason buyers default to one company over a better one. Category gravity builds slowly and it drains slowly, so on any given day it looks fine. The only way to see it is to measure direction over time, and it goes unmeasured.
What an annual relevance audit is
It is a structured read of how the market currently sees you, done once a year during planning season, before next year’s strategy gets written on top of it.
At Petrichor we run it as the Reality Audit. We score five dimensions from 1 to 5, and each one is scored against three reference points.
The first is what buyers say. We look at how they describe their own problem in their own words, in sales calls, in RFPs, and in the reasons they give when they pick someone else. The second is what AI engines cite. When your buyer asks an assistant for a shortlist, we check who gets named and what the engine points to as evidence. You don’t need to understand how that works. You need to know whether you’re in the answer. The third is where the category is moving, meaning which competitor is reframing the problem and whether the definition of the obvious choice is shifting under you.
The work takes about 12 days from kickoff to the readout.
The trendline is the product
A single audit is useful, but it isn’t the point.
Take the composite company again. Suppose its first audit gave it a 3 on how buyers describe the problem. A 3 on its own tells you very little. If that dimension was a 4 last October, the score means a competitor is pulling the definition away from you and you have a window to respond. If it was a 2 last October, the score means last year’s fix is working and you should keep going. The same number leads to opposite decisions, and you can only tell which one you’re in by comparing it to the year before.
That comparison is what the second and third audits are for. Which dimension slipped. How fast it slipped. Whether the fixes you funded last year actually held. A company that has three years of scores can see its own category gravity rising or falling long before the revenue line moves, and it can put budget where the slope is steepest instead of where the loudest executive is pointing.
This is the front door to the full Reputation Engineering engagement, and it is meant to be one. Sometimes the audit shows a slip too large for a planning-cycle fix, and that is the work reputation engineering exists to do. Just as often, the honest readout is that the position is holding and the plan can stand. Either answer is worth having before the budget locks, and both are cheaper than discovering the slip from your pipeline.
What it means for you
If you own next year’s strategy and budget, the practical move is small. Put the audit on the calendar for October or early November, ahead of the offsite, so the plan is built on a current read and not a remembered one. Then ask three questions of whatever you’re about to approve. Do buyers still describe the problem the way our plan assumes they do? If a buyer asks an AI assistant who to shortlist, are we in the answer? And has any competitor changed what counts as the obvious choice since the last time we checked?
If you can’t answer those from evidence, the plan is resting on a guess about the market, and it is a guess you can price out before you spend against it.
You wouldn’t write next year’s budget from last year’s books. Don’t write next year’s strategy from last year’s read of the market.