CaseStudies.Chat
← Back to the archive
change.archi2012–2021 · strategy

Sonalight's founders killed their app to build $1.35B analytics firm Amplitude (interview)

In a CNBC interview, CEO Spenser Skates describes shutting Sonalight (500K downloads, weak retention) to build Amplitude, now a $1.35B public company.

What was changed

In an interview with CNBC Make It published 18 September 2023, Amplitude co-founder and CEO Spenser Skates recounted shutting down Sonalight, the voice-recognition startup he built with Curtis Liu. The 2011 Y Combinator company — 'a version of Siri before Siri even existed', he says — earned press, seed funding and 500,000 downloads. But internally the founders watched people use the app once and never re-engage.

'Sonalight was a 95th percentile idea,' Skates said in the interview. 'It was pretty good, but it's not the bestest best.' Rather than grind toward 'some OK success', the pair spent a month exploring ideas and settled on the analytics tools they had built in-house to understand Sonalight's own users — 'we probably spent half of our time doing that', he says. They talked to 30 companies, found enough demand for the same insights, and began Amplitude in 2012, launching in 2014 with Jeffrey Wang as third co-founder.

Voice recognition had been 'almost too hard technically to solve', Skates recalls — a probabilistic problem without a clear answer — while analytics 'was a cakewalk' for two MIT algorithms engineers. Per CNBC's profile, Amplitude raised $336 million by 2021, when Skates took the company public, and stood at a $1.35 billion market cap with more than 2,300 corporate clients as of September 2023.

Why it worked

Usage data showed one-and-done behaviour; the founders read it as evidence Sonalight would plateau at 'OK success' rather than break out.

Voice recognition was technically intractable for a two-person team, while analytics played directly to their strengths as algorithms engineers.

They validated demand before building, talking to 30 companies and finding several that wanted the exact insights their in-house tool produced.

Skates says most engineers avoid startup risk entirely; the founders chose to absorb the uncertainty of abandoning a working product.

What can be applied

A 'pretty good' product can be the biggest trap: retention data, not press or download counts, tells you when to kill it — and the pivot often hides in the tooling you built for the first idea.

Aftermath

Amplitude launched in 2014, raised $336 million and went public in 2021; CNBC reported a $1.35 billion market cap and more than 2,300 corporate clients as of September 2023. Sonalight is gone, and the shutdown is now told as the deciding moment of Skates' career.

Sources

  1. 35-year-old CEO dumped his 'pretty good' startup — and built a $1.4 billion business instead ↗