There is an enduring temptation for startups that build consumer applications: Instead of creating a product for the end user, they let organizations with money influence their product.

It’s risky. Take it from a guy who knows.

Logan Green, CEO of on-demand ride-sharing company Lyft, spoke during a fireside chat at the Play Conference 2013, an event organized by students from the Haas School of Business at UC Berkeley. The conference drew a crowd of about 600 people, including students, entrepreneurs, and recruiters. 

On the topic of temptation, Green said he learned a big lesson from the early days of his first company, Zimride, a ride-sharing service for college students and employees of local companies. Launched in 2007, the company made money through fees it charged universities and companies for this service.

After it launched, Green did what any good entrepreneur would do. He listened to customer feedback. He soon realized there was a problem.

“Every transportation department at schools thought that there should be a bike buddy program. They thought that bicycle commuters wanted to find another buddy to bike with to campus, which is a nice idea,” Green said. “But the consumer demand for that was about zero.” 

But because the schools were paying clients, Green found himself between a rock and a hard place.

“[The company's] entire roadmap was kind of hijacked by the fact that we needed to sell to these people,” Green said. 

In 2012, Green and co-founder John Zimmer launched the mobile app Lyft with a new approach.

It allows peer-to-peer ride-sharing using a donation model. It also includes public ratings for both riders and drivers. For each ride, the company takes 20 percent of the sale, and 80 percent goes to the driver. Today, the San Francisco-based company has $82 million in total funding and reportedly facilitates 30,000 rides a week. 

The lesson: don’t let big customers knock you off course just because they have money.