How Yelp adapted to what users did instead
Yelp's first product asked friends for tips by email. What its users actually wanted to do was write reviews, and the company noticed in time.
You have to be very nimble and very open minded. Your success is going to be very dependent on how you adapt.
The version of this line that usually circulates reads “how your adapt”, a slip in transcription corrected here.
Jeremy Stoppelman and Russel Simmons started Yelp in San Francisco in 2004, both having worked at PayPal. The first product was built around email: a user would ask friends for a recommendation, say for a dentist, and the friends would reply through the site. It did not catch on. Few people wanted to send those requests, and fewer wanted to answer them.
Buried in the product was a smaller feature that let people write a review without being asked. The founders noticed that some users were writing long, opinionated reviews of restaurants and stores on their own, for nobody in particular. They rebuilt the service around that behavior, and the reviews became Yelp. Late in 2009 the company was in acquisition talks with Google that did not lead to a deal, and it has kept growing on its own since.
The story shows what the line means in practice. Being open-minded meant accepting evidence that the founders’ main idea was weaker than a side feature. Being nimble meant rebuilding quickly, before the money ran out. Plenty of teams see the same kind of signal and explain it away, because the main idea is the one they pitched to investors.
Adaptation has costs that the line does not mention. A company that adapts to every signal ends up with no shape at all. Yelp’s own later growth depended on something that was hard to change quickly: a large sales team selling advertising to local businesses, with contracts, habits and expectations. Once a company has made commitments like that, nimbleness becomes expensive, and adapting too readily can break promises to customers and staff.
The most useful form of the advice is about the early period, before those commitments exist. That is when users’ actual behavior is the best information a founder has, and when changing course costs the least.