Amplitude

Spencer Skates spent a year before launching any company reading about founders and startups, studying how successful companies were built. He read "Founders at Work" by Jessica Livingston and immersed himself in founder stories, trying to understand if entrepreneurship was right for him. After deliberate reflection on what would create the most impact in his life, he convinced his co-founder Curtis Yan to leave his job at Google in 2011 and start building together.

Their first real product was Sonalite, a voice recognition app that let users send and receive text messages by talking to their phone. Launched in summer 2011 around the same time as Siri, it gained thousands of downloads and got them into Y Combinator's winter 2012 batch. Their demo day presentation was slick—Spencer would put his phone in his jacket pocket and have a conversation with it, impressing the entire room. They got the most press mentions of any company in their 60-startup batch.

But then reality hit hard. Despite the press and downloads, users wouldn't retain in the app. Within weeks, their user base began shrinking. Spencer faced a dark moment of self-doubt: "I thought, you know, we're on this exponential curve and like, you know, it's no longer happening. Like what's going on here? I was going to be a failure." As engineers, they dug into the data and discovered the core problem: voice recognition accuracy was poor, and users would accumulate failed recognition attempts and simply stop using the app.

The failure of Sonalite became the seed for Amplitude. While building analytics to understand their retention problem, people they showed it to expressed interest—not in the voice tech, but in the analytics itself. They realized other product teams had the same pain: understanding how users actually engage with their products.

When they started Amplitude, they made a deliberate choice to flip their approach from Sonalite. Instead of building first, they spent their first month talking to 30 potential customers without building anything. They discovered that existing tools like Flurry had real problems: inaccurate data, lack of real-time reporting, and limited insights.

For about a year, they alternated between talking to customers and building what customers asked for—always giving it to them for free to get feedback. The big mistake Spencer identifies: they didn't ask for money early enough. "You don't have to ask for much, but just some amount of commitment from their end...even like 50 or $100 a month would have been a really big deal." This led them to spend time building for people who didn't truly value the solution.

Despite this misstep, the core motion worked: talk to customers, understand their problems, build solutions. As Spencer puts it: "all you got to do is build the product, sell it, build it, sell it, build it, sell it. Like that's it."

Amplitude launched four years before the interview and had grown to 100 employees. The company raised a Series C of $30 million about five or six months before the interview. They now have tens of thousands of products sending them data and hundreds of enterprise customers. Spencer's key insight on competition: most analytics companies compete on being "the most scalable and easiest to use." Amplitude carved out a specific niche—product analytics for product teams—rather than trying to be a generalist analytics tool. As he explains using Slack as an example, there are thousands of messaging tools but Slack dominates by focusing specifically on team collaboration. Amplitude does the same for product teams trying to build better products.