Product market fit for small products: the signals we watch

Quick answer: product market fit means people want what you have built enough to find it, pay for it and keep using it without you pushing. For a small product we do not wait for one big number. We watch a few plain signals in how customers arrive, return, ask for help and pay, and review them on a fixed date.

At Unavoidable Studio we build and run several products at once, including Unavoidable CRM, AisleReply, Unavoidably Secure and Unavoidably Social. Each one started with a validated problem, as described in how we validate an idea before we build it. Validation tells us a problem is real. Product market fit tells us our answer to it is working. This article is about the second part.

What is product market fit?

Product market fit is the point where a product clearly satisfies a real market. The phrase is usually traced to Marc Andreessen, who wrote in his 2007 essay The only thing that matters that “product/market fit means being in a good market with a product that can satisfy that market.”

We find it more useful to think of it as a change in effort. Before fit, every customer feels pushed: you explain, you remind, you chase. After fit, customers start pulling: they find you, they set themselves up and they ask for more. Our job is to notice which way the effort is flowing.

Why is product market fit different for a small product?

A small product does not have enough customers for statistics to settle the question, so you have to read behaviour instead. When a product serves a focused group, such as UK trades businesses or couples planning a wedding, a dashboard full of percentages can swing wildly from one week to the next.

So we look at what individual customers do, one by one, and we look for the same pattern repeating. One enthusiastic customer is an anecdote. The same behaviour from customer after customer, in the same kind of business, is a signal.

Which signals do we watch?

We watch six signals, in roughly this order of weight:

1. People arrive without being chased. Sign-ups that come from search, a recommendation or a partner, rather than from anything we sent, show that the problem is being searched for and our answer is being found.
2. People come back on their own. Customers who return week after week without a reminder email are using the product as part of their routine, not as an experiment.
3. People pay when the trial ends. Unavoidable CRM, for example, has a 14-day free trial. Someone adding a card at the end of it is the clearest yes we can get.
4. Support questions get deeper. Early questions are about getting started. When questions shift to edge cases and “can it also do this?”, people have moved in.
5. People notice when something breaks. If a small fault gets reported within minutes, the product matters to someone’s day.
6. The same type of customer keeps appearing. When sign-ups cluster around one kind of business or one situation, we know who the product is really for, and we can write and build for them.

Which signals do we treat with caution?

We treat compliments, launch-day spikes and loud feature requests with caution, because they feel like progress without proving it. In particular:

• Praise without use. “This looks great” from someone who never logs in again tells us the idea is appealing, not that the product works.
• Traffic spikes. A burst of visitors after a mention is welcome, but we judge it by how many of those people are still around a month later.
• Sign-ups that never set up. An account with nothing in it usually means the first ten minutes need work, not that demand is there.
• One loud request. A feature asked for by one customer gets noted. A feature asked for by several customers of the same type gets built.

What do we do when the signals are weak?

When the signals are weak, we narrow the audience, fix the first experience and set a date to decide. We do not keep adding features in the hope that one of them unlocks demand. Our usual steps are:

1. Talk to the people who left. A short, honest conversation with someone who stopped using the product is worth more than a survey of people who stayed.
2. Narrow who it is for. If one type of customer behaves differently from the rest, we focus the product and the words on them.
3. Fix the first ten minutes. Most drop-off happens before people see the product working, so setup is often the cheapest thing to improve.
4. Set a review date. We agree in advance when we will look again and what would count as better. If nothing has moved by then, we change direction or stop.

Stopping is a normal outcome in a studio. It frees the team for the next idea, chosen the way we describe in how we choose what to build next.

How do we track it across several products?

We track product market fit for every product with the same short checklist, reviewed on a regular date, and we keep each product’s results separate. The checklist is the six signals above, each marked stronger, the same or weaker since the last review, with a line of evidence next to it.

Using one checklist across the portfolio stops us from judging a product we love more kindly than one we are unsure about. Keeping the results separate matters too: each product is its own business with its own customers, so we never blend them into a single studio figure. That same one-team, separate-products approach runs through how we work, as set out in running several products with a small team. We share the process openly, as explained in why we build in public, and keep the figures for each company private until they are ready to share.

For the full list of products, visit the portfolio, and for more from the studio, browse the blog.