Your product fits today. In twelve months, perhaps not.

Two years ago, Product-Market Fit was an achieved state. Whoever had it, had made it. Today it's a lease with a short notice period.

The metrics that show it early are not in revenue. They show in renewal rates falling below last year, in sales cycles lengthening by two to three months, in net revenue retention slipping below 100 percent, and in churn reasons where "we'll do it ourselves now" suddenly appears.

This isn't happening because you got worse. It's happening because what you sell can be rebuilt differently than three years ago. An integration that took three engineers six weeks is now a three-day task. A contract review that took a senior manager a day now takes an hour. What cost days now costs hours. What cost hours now costs minutes.

In this world, the product is no longer your defence. It's your starting line.

Where this becomes visible operationally

The pattern usually shows up in the same quarter. A competitor doesn't appear from the direct market but from an adjacent one. A client tests an in-house solution that's seventy percent good enough and cuts the contract by two thirds. An industry where you knew six providers suddenly has twelve, six of which didn't exist two years ago.

The reflex in many leadership teams is: build faster. More features, more versions, more pace. That's the answer that reproduces the problem. When building gets cheap, more building isn't an advantage. It's noise.

The right question is different: which layer of my business is hard to reproduce today? If the answer is "the product", it's already too late.

Three layers, one diagnosis

The business has three layers.

At the bottom, the product. It used to be the defence. It's now the entry fee.

In the middle, the signal: the observed behaviour of customers. What they renew, what they pay for, who they recommend it to. Not what they say in the survey. What actually happens on Thursday afternoon.

On top, the fit: the repeatable pattern in which the market pulls toward the product. Not a single metric, but a constellation of stickiness, referral and willingness to pay.

Defence shifts upward. What can be protected is no longer protected by what is built, but by what the customer base does with you.

Five anchors that hold today

From work with leadership teams of mid-size firms, five anchors have proven robust. Not all apply to every company. But each company should be able to honestly name at least two.

First: proprietary data. Not the model, but what it has seen. Claim data over ten years, supplier behaviour, complaint patterns. Data a competitor cannot replicate in 24 months because it was created in reality, not on the web.

Second: industry knowledge held by the company. Not in two heads. Not in a slide collection. But in processes, decision rules, escalation paths. Knowledge that stays when the key person quits tomorrow.

Third: distribution relationships. Concretely: distribution contracts, purchasing decisions that have been falling your way for years because trust was earned. These can neither be built in five weeks nor replaced by a better tool.

Fourth: market position and brand trust. Whoever is known in a niche as the address you turn to when something must not go wrong has an advantage that AI doesn't halve. Trust is built slowly and destroyed quickly. It's asymmetric to market speed.

Fifth: protective rights and exclusive methods. Patents, regulatory accreditations, procedures only you are certified to apply. The thinnest anchor for most, the most important for some.

If you check these five anchors against your business and none holds, your current PMF is a lease with a notice period you don't know.

What to do operationally now

First: the diagnosis on a single page. Which anchor carries your business today? If you can't say it in five sentences, it's not clear enough to invest behind.

Second: one hard metric per quarter. A cohort, a renewal rate, a willingness to pay. Not a strategy update. A number whose movement you see before it shows in the P&L. When the fit tips, it tips here first.

Third: AI as leverage on layer one, not layer two. Speed up building, writing, prototyping. But don't fool yourself by manufacturing artificial signals. Layer two has to come from people who put money, time or reputation on the line. Otherwise it isn't learning, it's reassurance.

The uncomfortable consequence

Most leadership teams that have found their PMF in recent years treat it as achieved. That was never quite right. Today it's operationally expensive.

The contract with the market is running. It doesn't need to be renegotiated daily. But it needs to be reviewed every quarter. The firms that are still in fit two years from now are not the ones with the prettiest product. They are the ones that took their anchors seriously in time.

If you cannot honestly name an anchor today, that's the most important strategic question of the next half year. Not the next feature.