Product-Market Fit = Retention
There is one metric for product-market fit: do customers come back?
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 4
- Confidence
- 95%
Levine argues product-market fit means you're creating value, and it collapses to a single metric — retention. If people return, you create value; if they don't, you don't, and the company dies. He also notes that once a company finds PMF, the product essentially stops changing.
Origin
Uri Levine, drawn from the Waze global launch, where the product churned everywhere except four countries until iterated to 'good enough.'
Core principles
- 01If you don't figure out product-market fit, you die — no exceptions.
- 02Retention is the one metric that proves value creation.
- 03Once a company finds PMF, users use the product the same way as the first time — the product stops changing.
- 04PMF takes years and is the longest phase of the journey.
How to run it
- 1
Ship a not-good-enough version
Accept that you start below 'good enough' and that perfect is the enemy of good enough.
Pro tip 'The biggest enemy of good enough is perfect' — aim to be good enough to win the market, then iterate.
- 2
Measure retention against frequency of use
Track whether and how fast users come back. For high-frequency products, a 3-month retention of ~34–50% is a good indication.
Pro tip Conversion typically happens by the third or fourth use; if they don't reach the fourth time, they've quietly churned.
Watch out Retention thresholds depend entirely on natural frequency of use — a once-a-year product needs a different lens.
- 3
Interview churned users and rebuild
Talk to users who left, ask what didn't work, and build the next version addressing everything they told you.
Pro tip Users who wanted you to succeed will tell you exactly what failed — use it.
Watch out Expect to be wrong repeatedly; conviction that 'this is it' is often premature.
- 4
Iterate until good enough, then stop changing the product
Repeat the churn-interview-rebuild loop until retention holds, then shift the company's focus to growth or business model.
Watch out Trying to work on growth or business model before PMF splits focus and stalls the company.
In the wild
After success in Israel, Waze went global end of 2009 and was 'not good enough' almost everywhere except Czech Republic, Slovakia, Latvia, and Ecuador. Over a year of churn-interview-rebuild iterations followed until it worked in early 2011, then spread metro by metro and country by country.
→ Reached good-enough retention market by market; became a category-defining navigation app sold for ~$1B.
Levine points out we search Google, use Waze, WhatsApp, Netflix the same way as the first time — because once PMF is found, the product stops changing.
→ Illustrates that PMF is the point at which value is locked and product churn ends.
Common mistakes
Crowdsourcing a problem that needs perfect data
In Japan, house numbering is chronological (oldest house is #1), so approximation fails — crowdsourcing can't reach 'good enough' where perfect information is required.
Chasing vanity metrics over retention
Downloads and sign-ups mean nothing if users don't return; retention is the only proof of value.
Is it for you?
Best for
Founders trying to know objectively whether they have product-market fit.
Not ideal for
Very low-frequency products (e.g. annual tax filing) where retention is a weak near-term signal.
From the transcript
“product Market feed have one metric one metric that's it retention”
“if you create value they will come back if they are not coming back that means that you are not creating value”
“the biggest enemy of good enough is perfect”
From the episode
Lessons from a two-time unicorn builder, 50-time startup advisor, and 20-time company board member
Uri Levine (co-founder of Waze)