How a product is born. Part 2: Product-market fit

In the previous parts, we covered the idea and the MVP. Now the product is live. Users may show up, or they may stay away. This is where things get interesting.

How do you know whether the market needs the product? How do you distinguish passing interest from sustained demand? That is what product-market fit, or PMF, helps you understand.

PMF is the fit between a product and market needs. Observable signals help assess it: how people use the product, whether they return, and whether they are willing to pay.

Three problems most teams encounter

After launch, teams tend to encounter three recurring problems. Each brings a different question.

First, there are no sign-ups, users or activity. You launched, and nobody came. That alone is hardly a catastrophe. It does suggest that some work was left unfinished before launch.

Second, people sign up but never return after their first visit, or barely do anything inside the product. You have failed to engage them. They registered, tried the service and disappeared. You need to measure how often that happens in your own product.

In fintech, this problem can be particularly acute because of KYC and other regulatory requirements. More complex onboarding means more opportunities for people to drop out.

The third problem requires you to get further, which is an achievement in itself. People register, return and use the product, but they do not pay. The question becomes: how do you turn an active audience into paying customers?

You can work on all three problems when you have data. Without data, you are working in the dark.

What needs to be ready before and immediately after launch

In practice, some startups launch without basic analytics. The team cannot tell who arrived, where they came from or what they did inside the product.

Here is the minimum you should prepare.

Analytics and UTM tags

You need to understand where each user comes from. Otherwise, you cannot expand the channels that work or stop the ones wasting your budget.

And yes, you need a budget. A product will hardly develop on its own simply because it is good or people supposedly need it.

UTM tags are a basic tool, but they help answer a useful question: which channel brings you the right people?

Onboarding analytics

You need to see where people abandon registration. Which field did they leave empty? Where did they stop? What put them off?

There is a balance to strike with barriers. Too many, and conversion falls because people never finish. Too few, especially in fintech, and the product can look suspicious. A careful user may leave for that reason too.

Start with the customer's needs. What do they need to do at this step to feel confident and reach a useful result?

Notifications

Someone may have registered yesterday and forgotten about you today. They had a problem, found your product, started doing something and left.

A notification can bring them back at a useful moment. In fintech, transaction notifications help people track their funds, notice problems and trust the service. Which notifications are mandatory depends on the product and the requirements that apply to it.

A working core journey

On launch day, the main journey delivering the product's value and revenue needs to work reliably.

One clear path that takes the user to a result is enough to start. Sixteen different monetization methods can wait.

Marketing needs a budget

One budget mistake might look like this: 80% goes into development and 5% into marketing. That can leave too little money to find customers.

In an offline business, you can open in a good location and expect some people to walk in. Software rarely works that way. Building a product hardly guarantees that anyone will hear about it.

Customers do not arrive simply because the platform is open.

For comparison, consider a different allocation: around 30% for development, 10% for the internal team and the rest for marketing. This illustrates a budget structure. The proportions that work depend on the project.

You still need to pace that marketing spending. When you spend faster than you learn why something works or fails, you burn money without gaining knowledge.

Start with small tests. Spend less than you can afford. Look for a combination of audience, message and channel that works. Watch who arrives, why they stay and why they leave.

A typical situation: registrations jump, but purchases do not. Your campaign may be attracting the wrong people. You need to review either your audience or the product itself.

For financial planning, test a scenario in which the first paying customers take 4 - 6 months to arrive in B2B or fintech, or 2 - 3 months in B2C. These are assumptions for calculating your cash buffer; estimate your own timeline separately. I would plan for at least six months of marketing after launch.

The first 100 customers take hands-on work

For many products, the first 100 customers come through conversations, direct outreach and the founder's personal involvement. An automated funnel comes later.

In Paul Graham's essay, Airbnb worked with hosts in person, while Stripe's founders set up early users themselves. That work helps get a product started and reveals how people use it.

Graham called this doing things that do not scale.

Early on, some processes are impossible or too expensive to automate. Sometimes the sensible choice is to do them yourself or ask someone on the team to help.

You gain customers and an understanding of your audience. You see who buys, why they buy, what matters to them and where they hesitate. Customer interviews alone cannot fully replace that experience.

Places to look for your first customers include:

• Relevant communities and group chats, which exist in almost every niche.
• Direct outreach on social platforms.
• Partnerships with relevant creators who already reach your audience.

Those partnerships can take time to negotiate. A good partner can bring a valuable stream of users with a high NPS. We will return to that metric shortly.

The main post-launch mistake: scaling a problem

People register but never stay, and the team increases its advertising budget.

Registrations grow, nobody knows where they come from, and spending increases anyway.

The audience is large but only a handful pay, and the team raises investment to grow it even faster.

In each case, the team scales an error.

The numbers can look like growth. Meanwhile, resources, sometimes already raised from investors, go into expanding the problem.

Here is a case from my own practice. A product attracted more than 100,000 users, but only six paid during the three months after a paid version was introduced. Dear reader, you read that correctly. The founder scaled their own idea of the product. The market had a different view.

The team had to pivot and turn the platform into an information portal. It already had an audience, and advertisers were found quickly. That time, things worked out. They could have gone another way.

The aha moment: when a user understands the value

The aha moment is the point at which someone realizes they need the product.

Curiosity turns into a desire to keep using it.

In a 2015 interview with Slack's founder, that signal was 2,000 messages sent within a team. Of the teams that reached that milestone, 93% were still using Slack. The figure described continued use, without claiming that those teams renewed a paid subscription.

On a social platform, the aha moment might arrive when someone follows 30 accounts they find interesting and feels they have found a comfortable place.

Finding this moment in your own product is difficult. It usually emerges after launch, mainly through behavioral data and sometimes through conversations with people who stayed.

Once you understand it, you can redesign onboarding to guide people toward that point.

Understanding the aha moment helps remove unnecessary steps and get users to the experience that makes the product valuable to them.

Metrics you need to measure

The following metrics help you understand where users find value and where you lose them.

Retention

Retention measures whether users return.

A hundred returning users can matter more than 10,000 registrations. Signing up is the first step. Value becomes visible when someone comes back.

Compare industry benchmarks only when the time period and definition of activity match. Daily messaging activity and a repeat purchase in an online store describe different behavior. Decide which action and period count as a return, then compare your cohorts.

If retention declines or falls behind comparable cohorts, investigate.

When registrations grow without returning users, find out why people came and what they could not find. Contact people who registered and left. Offer something in return for an interview.

You will often hear something like, “I couldn't find this,” or, “I didn't understand what was happening here.” The second answer is less pleasant, but useful too.

Find out what someone expected to see and why they went through registration. The expectation that brought them in, and remained unmet, may define your next version of the product.

Cohort analysis

Cohort analysis shows how retention changes over time for different groups of users.

For example, January's cohort retains at one level, February's lower, and March's lower still, although the product barely changed. Marketing may have started bringing in the wrong audience.

Cohort analysis helps you notice that before it makes a visible dent in revenue.

Churn rate

Customer churn is the share of paying customers who stop paying during a given period.

Check it regularly. A weekly view can be useful for products with frequent payments. Your comparison period needs to match the payment cycle.

A product can grow in registrations, retention and new payments while losing existing paying customers. Aggregate figures may hide that, and losing established customers is expensive.

When the number of paying customers falls, speak to those who left.

They may have found a competitor with better terms, become dissatisfied with the service or faced changes in their own business. Each reason gives you something to fix or useful information for your strategy.

NPS and the Sean Ellis test

Net Promoter Score starts with a question: how likely are you to recommend this product?

Getting useful answers is harder than asking the question. Collect scores in a way users understand, then explore their reasons in interviews. The survey method itself can affect the result.

For an initial working view, you might aim for 50 - 100 responses. Treat small samples carefully; a particular sample size alone never guarantees accuracy.

Users arriving through a trusted creator may give higher scores than cold traffic. Account for the acquisition channel when interpreting the results.

Under the Bain methodology, NPS is the percentage scoring 9 - 10 minus the percentage scoring 0 - 6. Scores of 7 - 8 remain in the total sample but add nothing separately to the calculation. The share of high scores is useful, although it cannot establish PMF on its own.

The Sean Ellis test asks a different question: how would you feel if you could no longer use this product?

In Ellis's original methodology, the benchmark is more than 40% answering “very disappointed.” That is a strong value signal for the audience surveyed. Ask people who recently experienced the product's core offering. Combining all negative reactions would give a different measure.

What PMF means in the end

Product-market fit is assessed through a combination of signals. They show how well the product is holding its place in the market after launch.

Much of that evidence comes through feedback: conversations, interviews, onboarding data and behavioral analysis.

Talking to users is therefore an essential part of early product work.

Consider this scenario. You enter a market first. The product is still weak, but there are no competitors, so you grow because you arrived first. You teach the audience a new way of doing things.

Then a competitor arrives. They listen to feedback, improve the experience and start winning your users.

You created the market and taught people how to use the new solution, only to become its second or third player.

Failing to measure delays the discovery of problems.

You learn about them later. And almost always at a higher cost.