What to Measure Before Product-Market Fit

Most startup metrics advice is wrong for pre-PMF companies.

MRR growth. CAC payback. Net revenue retention. Funnel conversion rates. These metrics matter—later. Before product-market fit, they often mislead more than they inform.

The problem: growth metrics assume you know what works. They help you optimize. But if you're still searching for product-market fit, you're not optimizing—you're exploring. Different activity requires different measurement.

Why Growth Metrics Mislead Early On

Growth metrics aggregate behavior across customers. They work when your customers are relatively homogeneous—when you've found a segment that responds consistently to your product.

Before PMF, you don't have that consistency. Your users might include:

• Early adopters who love the product
• Curious browsers who'll never return
• People from the wrong segment entirely
• Friends and family giving biased feedback

Aggregate metrics blend these groups together. Your "retention rate" combines passionate users with accidental signups. Your "conversion rate" mixes serious buyers with tire-kickers.

The signal drowns in noise.

The Trap of Premature Measurement

Some founders respond to this by building elaborate analytics infrastructure early. Dashboards. Funnels. Cohort analysis. All tracking users who number in the dozens.

This creates several problems.

False precision. When you have 30 users, a 20% retention rate means 6 people stayed. Is that good? You don't know. Statistical significance requires volume you don't have. Metric fixation. Once you're tracking something, you start optimizing for it. But early on, the metrics you can measure aren't necessarily the metrics that matter. Activity substitution. Building dashboards feels productive. It's easier than the uncomfortable work of talking to customers. But dashboards don't tell you why users behave as they do. Premature optimization. If your conversion rate is 2%, the temptation is to optimize the funnel. But maybe the funnel isn't the problem—maybe the product is wrong for this audience entirely.

What Actually Matters Pre-PMF

Before product-market fit, you're trying to answer a fundamental question: have you built something people genuinely want?

This is better answered through qualitative observation than quantitative measurement.

Behavioral intensity. Do users engage deeply, or just browse? Deep engagement with ten users is more promising than surface engagement with a hundred. How often do they return? What do they do when they're there? Unprompted actions. Do users tell others without being asked? Do they reach out with feedback? Do they request features? Unprompted engagement reveals genuine interest, not just polite participation. Willingness to pay. Not "would you pay" in hypotheticals—actual payment. Do people exchange money for what you've built? This filters out casual interest. Specificity of feedback. Generic praise ("this is cool") signals politeness. Specific feedback ("this part doesn't work for my use case because...") signals engagement. The depth of feedback reveals the depth of caring. Disappointment at churn. When users stop using the product, do they disappear silently or do they explain why? Users who bother to tell you why they're leaving care more than users who ghost.

Signals Worth Tracking

Some things are worth tracking even pre-PMF—but with different framing.

Who comes back without prompting. Not a retention percentage—specific individuals. Who are they? What do they have in common? Why do they return? Where users get stuck. Not funnel drop-off rates—actual stories of confusion. What do users try to do that they can't? Where do they ask for help? What users say verbatim. Not NPS scores—actual quotes. The language users use reveals how they think about the problem and your solution. Who refers others. Not a referral coefficient—specific people making introductions. What's different about them? Why do they share? Patterns in feedback. What do multiple users mention unprompted? Convergent feedback suggests real issues or opportunities.

These aren't metrics in the traditional sense. They're observations. They require talking to users, reading support tickets, watching session recordings. They're qualitative, and that's appropriate for the phase.

The Conversation Problem

Metrics can substitute for conversations. That's dangerous early on.

If you track everything users do, you might think you understand them. But data shows what happened, not why. It shows behavior, not motivation. It shows patterns, not causes.

Early-stage startups need to understand why. Why did that user churn? Why did this one become a power user? Why do some people pay immediately while others never convert?

The only way to learn why is to ask. Customer discovery doesn't stop when you launch. It continues until you understand your users well enough that metrics can take over some of the work.

When to Shift to Growth Metrics

How do you know you're ready for traditional metrics?

Consistency appears. Your users start looking similar. The same type of person, with the same problem, behaving in similar ways. Patterns emerge without you forcing them. You can predict behavior. You know which users will retain before they do. You know which leads will convert. Your intuition, built from qualitative learning, starts being reliable. The numbers have volume. You have enough users that percentages mean something. A 10% change in retention represents real behavioral shifts, not random variation. You know what to optimize. You're not searching anymore. You've found something that works, and now you're making it work better. This is optimization, not exploration.

At this point, metrics become powerful. They help you scale what you've found. They identify leverage points for improvement. They enable growth.

But they serve optimization, not discovery. Use them in the right phase.

The Real Pre-PMF Metric

If forced to choose one metric for pre-PMF, it might be this: how many customers would be genuinely disappointed if your product disappeared?

This gets at the core of product-market fit. Not satisfied customers—disappointed ones. Not users who'd find an alternative—users who'd feel a loss.

You can measure this with the Sean Ellis test: ask users how they'd feel if they could no longer use the product. Count the "very disappointed" responses.

But even this is a proxy. The deeper question: are you building something that matters to someone? That question gets answered through depth of engagement, not breadth of measurement.

Related Reading

• Signs of Product-Market Fit
• The Metrics Theater: Tracking Numbers That Don't Matter
• Sean Ellis Test Explained
• Product-Market Fit Metrics

Unsure what to measure at your stage? Take our free PMF assessment to understand where you are and what signals matter most right now.