Product-market fit is not a feeling you get on a good day. It is a verdict you reach by reading evidence carefully. For most first-time founders the danger is not a lack of data, it is the temptation to read your own data to confirm what you already hope is true. This lesson gives you a disciplined way to judge whether you are genuinely approaching fit or quietly fooling yourself.
Separate vanity metrics from real evidence
Downloads, signups, page views, press coverage, a funding round, and even DPIIT recognition all feel like progress, but none of them proves that people need your product. Eric Ries calls these vanity metrics in The Lean Startup: numbers that only ever go up and flatter you. Evidence answers a harder question: are the same users coming back, and paying, without you prompting them?
Be especially careful in the Indian market, where discount coupons, cashback, festival-season spikes, and free trials can inflate usage that collapses the moment you stop subsidising it. A number you paid for is not a number you earned.
The retention curve is your honesty test
Group your users by the week or month they first signed up, which is called a cohort, and track how many stay active over the following weeks. If each cohort's curve keeps sliding toward zero, users are not finding lasting value. If the curve flattens into a stable plateau, a real segment has adopted you. That plateau, not a launch-day spike, is the clearest quantitative sign that you are nearing fit. Compare cohorts over time: if newer cohorts flatten higher, or flatten sooner, your recent changes are working.
Ask the one question that predicts fit
Growth expert Sean Ellis popularised a simple survey: "How would you feel if you could no longer use this product?" with three options, very disappointed, somewhat disappointed, and not disappointed. As a practical benchmark, 40 percent or more choosing "very disappointed" signals product-market fit. Below roughly 25 percent, you have not found your market yet. Between 25 and 40 percent, build hard for the people who said "very disappointed."
The honesty rule matters more than the number: survey only users who have actually experienced the core of your product, for example those who used it more than once recently, not everyone who ever created an account. When Rahul Vohra's team at Superhuman surveyed only engaged users and then split their roadmap between deepening what fans loved and removing what blocked fence-sitters, their score climbed from 22 percent to 58 percent. Averaging in casual tourists would have hidden the truth.
Net Promoter Score, the "how likely are you to recommend us" question introduced by Fred Reichheld of Bain and Company in a 2003 Harvard Business Review article, is useful for word of mouth, but it is a coarser, lagging signal. Do not mistake a decent NPS for proof of fit.
Read qualitative signals without flattering yourself
The strongest sign of fit is pull: users chasing you, sharing the product without being asked, getting genuinely upset when it breaks, and building their own workarounds to keep using it. The opposite is push, where growth only happens when you push through ads, personal chasing, or discounts.
Guard against two traps. Confirmation bias makes you listen to fans and ignore the people who left, so deliberately interview users who churned and ask why. Survivorship bias means your happy users are the least representative of the wider market. One warm message is a single data point, not a trend.
A weekly ritual to stay honest
- Name three vanity metrics you will stop celebrating.
- Pull one cohort retention chart and ask: is it flattening, or bleeding to zero?
- Track the "very disappointed" percentage among engaged users over time, not just once.
- Log one real churn reason and one moment of unprompted pull each week.
- Write your verdict in one honest sentence, then state what evidence would change your mind.
If you cannot say what would prove you wrong, you are not reading evidence. You are collecting reassurance.