Blog · 14 Apr 2026

Finding product-market fit before you spend on ads

The most expensive mistake we see is a founder with a working product and some funding deciding it is time to scale. Budget goes into ads, the numbers look busy, and six months later the money is gone and the retention curve is flat.

Marketing does not create product-market fit. It reveals whether you have it. Spending before you know is how founders buy an expensive answer to a question they could have answered for free.

Signals that come before revenue

Product-market fit is felt before it is measured. The early signals are behavioural:

• People come back without being prompted. Retention, not signups.
• Users get annoyed when the product breaks. Indifference is the real enemy.
• They describe it to other people in their own words, and those words are roughly right.
• Someone has hacked together a workaround because they need a feature badly.
• Sales conversations get shorter over time rather than longer.

The retention curve tells the truth

Cohort your users by the month they arrived and plot how many are still active over time. There are three shapes and only three:

• Down to zero. No fit. More traffic makes this worse, not better.
• Down then flat. A stable core exists. This is the shape you want. It means you have a real segment and the work is to find more of them.
• Down then rising. Rare and excellent. Existing users expand their usage.

If your curve goes to zero, no amount of ad spend fixes it. Every new user leaks out the same hole.

Talk to the people who left

Founders interview happy users because it is pleasant. The information is in the churned ones. Ask what they hired the product to do, what they use instead now, and what would have had to be true for them to stay. Twenty of those conversations will teach you more than any dashboard.

Narrow until it hurts

Most pre-fit products are trying to serve too many people. The fastest route to fit is usually subtraction: pick the single segment where the pain is sharpest, and build for them so specifically that a general competitor cannot follow.

"Inventory software for independent pharmacies with one to three branches" will beat "inventory software for retailers" every time at this stage, because you can learn faster, sell more cheaply, and be obviously the right answer for someone.

When you are allowed to spend

Turn on paid acquisition when three things are true: retention has flattened rather than gone to zero, you can state which segment the product is for without hedging, and you know roughly what a customer is worth over their lifetime. Until then, spend the money on conversations instead. They are cheaper and they compound.