How Uri Levine built two unicorns by falling in love with the problem

Executive overview

Most startups fail because founders chase solutions before validating a real problem. Uri Levine — co-founder of Waze, serial entrepreneur across 10 companies, and advisor to 50+ founders — argues that the problem must be the north star of the entire journey.

Fall in love with the problem first. Validate it with strangers. Only then build.

The single biggest predictor of startup success is whether you're solving a problem enough people care about — not whether your solution is clever.

Why the problem comes first

• Solve a problem you personally hate; frustration is the most reliable source of startup ideas
• Talking to ~20 people you don't know is enough to validate whether a problem is real
• Start your story with "the problem we are solving is…" not "our company is…" — the framing signals whether you're user-focused or solution-focused
• If the problem isn't the north star, every pivot becomes directionless

The four phases of building a startup

• Phase 1 — All over the place: thinking about everything; the only real job is figuring out product-market fit
• Phase 2 — Product-market fit: if you don't crack this, you die; the single metric is retention
• Phase 3a — Growth first (high-frequency products): word of mouth kicks in naturally; grow before monetising
• Phase 3b — Business model first (low-frequency products): without natural word of mouth, figure out how you'll make money before spending on growth
• Focus is not about what you're doing — it's about what you're not doing
• Shifting gears between phases is hard; companies that miss the gear change get stuck

Product-market fit in practice

• PMF = customers come back; retention is the only metric that matters
• Every major app you use today — Google, Waze, WhatsApp, Netflix — looks identical to the first time you used it; PMF locks the product
• Waze took four years; Microsoft five; Netflix ten; ChatGPT six years before anyone noticed
• The right retention benchmark for high-frequency products: 30–40–50% three-month retention is a healthy signal
• Watch for the third or fourth use — users who don't reach it have already decided it's not good enough
• Waze iterated for over a year across markets before PMF clicked; failing fast multiplied the number of attempts available

Fundraising tactics

• Investors decide if they like the founder before they sit down — first impression is formed in seconds
• Early-stage investors fund two things: "I like the CEO" and "I like the story" — that's it
• CEO should go to investor meetings alone; other people dilute the spotlight
• Expect 99 nos for every yes — the same 1% conversion rate a VC sees from 100–200 pitches per year
• Don't argue with a no; it's like arguing after being turned down on a date
• The most overlooked slide is the first one — it stays on screen the longest; put your strongest point there
• The second most overlooked slide is the last one ("Thank you") — repeat your strongest point; don't waste the airtime
• Good stories create emotional engagement, not just facts; details make stories believable
• Always start with the problem because investors are users too — if they don't feel the pain, they dismiss the market

Hiring and firing

• About half of failed startups cite the wrong team as the cause; most founders knew within the first month
• The real failure is not making the hard decision — top performers leave organisations that can't fire
• The 30-day calendar test: when you hire someone, set a reminder for 30 days and ask: "Knowing what I know today, would I hire this person?"
• If yes: tell them they're exceeding expectations and give them more equity — you buy loyalty for life
• If no: fire immediately; they are already on a trajectory of failure and deserve to find somewhere they can succeed
• To confirm a hard call, ask top performers: "How upset would you be if this person left?" — their answer is your signal
• For references, ask one question: "Would you hire them?" If yes, ask why they didn't

Understanding users

• Watch new users; if they're not doing what you expect, ask why — the "why" makes products better
• Different users adopt new behaviours at different rates: innovators (~2%), early adopters, early majority (~33%), late majority, laggards
• The early majority is the prize; they're afraid of change and won't try unless it feels safe and simple
• Product owners are usually innovators or early adopters — they literally cannot understand early-majority behaviour without observing it
• Talk to users who churned or never converted, not just the ones who succeeded — they know what's broken
• Waze users fell into two groups: 70% watched the screen, 30% listened to audio prompts; building only for one would have been wrong
• Simplicity is always the solution for the early majority; remove features people don't use — they add complexity without value

The startup mindset

• The journey has three dimensions: roller coaster (high frequency of ups and downs), journey of failures, and long duration
• Failing fast preserves more attempts; more attempts raise the probability of success — more shots at the basket
• The biggest enemy of good enough is perfect; iterate to good enough, then win the market
• "If you haven't failed, it's because you haven't tried new things" — Albert Einstein
• Ask yourself constantly: "Knowing what I know today, would I do this differently?" If yes, change it today
• Teach children to fail — when they get up, they get up stronger; they'll explore and find what makes them happy