Lenny's Podcast · Jun 9Lessons from a 2-time unicorn builder, 50-time startup advisor and 20-time board member | Uri Levine
Went global at the end of 2009 and worked in only about four places (Czech Republic, Slovakia, Latvia, Ecuador). It took over a year of rebuilding with churned drivers to work in early 2011, about four years to reach product-market fit. It raised about $50M total, sold for over $1B, planned to sell map data to governments but chose advertising, and never worked in Japan because of house numbering. 95 to 99 of 100 users say they heard of it from someone.
Tips
- Measure product-market fit by retention and watch the third or fourth use▶ 20:58Waze compared every new market to Israel by frequency of use and return rate.
- Iterate with churned users until the product is good enough▶ 23:00Waze took about four years to find product-market fit, and it worked in only about four countries at global launch (Czech Republic, Slovakia, Latvia, Ecuador).
- Focus on one startup phase at a time▶ 34:20Waze began by planning to sell map data and traffic info to governments, then moved to advertising because of long sales cycles.
- Pick growth-first or business-model-first by frequency of use▶ 35:20Asked how they heard about Waze, 95 to 99 of 100 people say someone told them.
- Start from the problem or user value, not your solution▶ 4:05In 2007 'an AI crowd-sourced navigation system' draws no interest, but 'I'll help you avoid traffic jams' does.
- Watch new users and ask why when they deviate▶ 1:07:20In Uri's talks about 70% of people use Waze by watching the screen and 20 to 30% by audio prompts; Wozniak told him he was the first person to take a selfie with the volume button as intended.
- Sell the value users actually get, such as certainty▶ 28:45There are more traffic jams today than in 2007, yet Waze still delivers value through predictable arrival times.
- Don't build crowdsourced products where you need perfect data▶ 26:00Waze worked in the US, UK and Israel but not in Japan.
