Pivot or persevere: how to tell the difference

A founder rang me in July, eight months into building something she believed in, and asked the question straight out. Should I pivot or persevere? She had users. Not many, but real ones. She had a product that worked. She also had a quiet, growing feeling that something was off, and no idea whether that feeling was a signal or just fatigue. We spent an hour on the call and did not once talk about vision or grit. We talked about what her numbers had actually been doing for six months. By the end she knew her answer, and it was not the one she had phoned up expecting.

Founders treat this as a character test, and it is not

The pivot or persevere decision gets framed as a test of nerve. The stories everybody repeats push you in both directions at once. Persevere and you are the founder who held on through the flat months until the thing finally worked. Pivot and you are the founder who was clear-eyed enough to kill their darling and find the real business. Both stories exist, both are true, and neither of them tells you which one you are in right now.

That framing does real damage, because it turns a question about evidence into a question about identity. Founders who think of themselves as resilient persevere past the point where the data has stopped supporting them. Founders who think of themselves as adaptable pivot away from something that was two months from working. In both cases the decision was made by self-image rather than by what was happening in the product.

I want to take the drama out of it. The question is not whether you have the strength to keep going or the humility to stop. The question is much narrower. Is there any evidence that the thing you built is getting better at doing its job, and if so, is it getting better fast enough to matter before you run out of money?

The three numbers that actually answer pivot or persevere

There is a version of this decision that is genuinely hard and a version that founders make hard by looking at the wrong things. Almost every founder I speak to is watching total signups. Total signups only ever go up, which makes them useless for this decision. They will reassure you right up until the day the bank balance stops agreeing with them.

The numbers that answer the pivot or persevere question are these three, in this order.

Retention at your natural usage interval. Take a cohort from three months ago and ask how many of them are still using the product at the rhythm the product implies. A weekly tool should be measured weekly, a monthly one monthly. If the curve flattens at any level above zero, you have found people with a real problem and a real habit. That flat portion is the only thing in your business worth building on. If the curve goes to zero every single time, no amount of marketing will fix it, because you are refilling a bucket with a hole in it.

The trend in that curve across cohorts. One flat retention curve is a starting point. The question is whether the cohort from last month retains better than the cohort from four months ago. That is the only honest measure of whether your work is compounding. If you have shipped twenty improvements and the curves are identical, you are busy rather than progressing, and that is the clearest pivot signal there is.

How hard you are having to push. Track the effort behind each new user. Founders routinely mistake their own heroics for product traction. If every signup came from you personally messaging someone, the product is not pulling anyone in, you are. That is fine in month one. In month eight it means the thing does not yet stand up on its own.

If you cannot answer these three, that is not a reason to guess. It is a reason to stop and instrument the product properly, which is a week of work. We wrote about setting up app analytics from day one precisely because founders keep reaching this decision with nothing to base it on.

What a pivot signal really looks like

Here is the part that surprises people. A genuine pivot signal is almost never an absence of interest. It is interest pointing somewhere you did not intend.

The clearest one is when users keep doing something adjacent to what you built. They sign up for the workflow tool and use it as a client tracker. They ignore four of your five features and hammer the fifth. They keep asking for an export, and it turns out the export is the product. That is not failure, it is a market telling you where the value sits, and the founders who hear it do very well.

The other real signal is quieter and worse. Nobody resists your product, nobody complains, and nobody comes back. Polite indifference is far more dangerous than criticism, because it produces no urgency. You can spend a year improving something that nobody dislikes and nobody needs.

What is not a pivot signal, despite feeling like one, is a bad month. Or a competitor raising a round. Or being tired. Or one loud user with strong opinions. I have watched founders throw away eighteen months of compounding work because a well-known company launched something adjacent, and in most of those cases the competitor was not the reason their retention was flat.

What perseverance should actually mean

Perseverance gets treated as a synonym for continuing, which is why it so often means doing the same thing more loudly. The useful version is narrower and it looks more like a series of controlled experiments than a display of endurance.

If you are going to persevere, name the specific belief you are persevering on. Not “people will want this” but something you can be wrong about inside six weeks, like “practice managers will pay £40 a month once they can import their existing patient list”. Then build the smallest possible test of that belief and set the date before you start. A founder who says “I am giving this eight more weeks and here is the number that decides it” is in a far stronger position than one who is simply refusing to quit.

This is also where the honest version of an MVP earns its keep. The point of building small is to keep the cost of being wrong low enough that you can afford to find out. If your next test takes four months and most of your runway, that is not perseverance, that is a bet, and you should name it as one.

What I would do differently

I have been on the wrong side of this myself. With an earlier product, I persevered for about five months past the point where the evidence had made its position clear. I knew the retention curve was going to zero. I looked at it most weeks. What I did instead of acting on it was ship features, because shipping felt like progress and deciding felt like loss. The tell, in hindsight, was that I had stopped showing the numbers to anyone else. Founders who are avoiding a decision go quiet about the metric that would force it.

So the rule I use now is embarrassingly simple. Pick the number that would change your mind, write it down, and tell one person outside the business what it is. Not a co-founder, who is inside the same story as you. Someone who will ask you about it in six weeks and not accept an answer about how busy you have been. Half the value of this decision is just making it in public rather than in your own head at two in the morning.

The other thing I would change is who I was listening to. In that five-month stretch I was talking almost exclusively to people who already liked the product. If you want a real read, the useful conversations are with the people who signed up and left. Five of those is usually enough to turn a vague unease into something concrete, which is the same reason we push founders to talk to five customers before building anything.

What to do next

If you are sitting in this decision right now, do these three things before you do anything dramatic. Pull your retention curve by cohort and look at whether it flattens. Compare your newest cohort to your oldest. Then write down the single number that, if it had not moved in eight weeks, would settle the question for you.

Most founders who do that find the decision has already been made and they were waiting for permission to see it. A few find genuine compounding that they had been too close to notice, and those founders should absolutely persevere, now with a reason rather than a hope.

If you want a second pair of eyes on the numbers before you commit either way, send us the short version of where you are. We have sat on both sides of this decision, including the times we got it wrong, and an hour with someone who has no stake in the outcome is usually worth more than another month of thinking about it alone.