Clerk's Hockey Stick growth, with Colin Sidoti
Clerk co-founder Colin Sidoti traces the accumulation of tiny product decisions behind the company's hockey-stick growth — and explains why the fastest route to value is more important than making the first component look impressive.
Colin Sidoti returns to the podcast as co-founder and CEO of Clerk, a user-management platform built around drop-in components for authentication and B2B organization management.
We discuss Clerk's years of iteration before reaching 7% weekly growth, framework-specific developer experience, testing improvements by watching people struggle, preserving founder context as a team grows, hiring technical developer marketers, and a van-based attempt to reach every Y Combinator company.
The Q&A below is experimental — extracted from the episode transcript with AI. If you spot a mistake, let us know at jack@scalingdevtools.com.
What is different about Clerk’s approach to authentication?
Clerk begins with React components for sign-up, sign-in and user profiles, with the service behind them. The component becomes the new API: OAuth, two-factor authentication and profile management can work without every team rebuilding the same flows. B2B organization components apply the same idea to invitations, roles and permissions.
How did the original product change at South Park Commons?
We began in 2019 believing Clerk would be a Ruby gem. South Park Commons continually brought in people starting new projects, so we placed prototypes in front of them, watched what resonated and moved toward React components. The pre-seed period bought the learning; the seed round then funded production-grade infrastructure.
What happened before the growth curve finally changed?
From the 2021 launch through early 2023, growth felt weak. We replaced an agnostic JavaScript pitch with deep framework-specific integrations, including Next.js, and redesigned session architecture after seeing repeated mistakes. None was a single magic feature. Together they made a reliable good first experience possible.
What did hockey-stick growth look like?
In early 2023 we first sustained the roughly 7% weekly growth benchmark I had carried from Paul Graham's writing. Then YouTube creators discovered Clerk and some weeks grew around 50%; later months produced very large revenue gains too. Even with those numbers, a founder's instinct is to believe the product can still improve.
What is the key moment in Clerk’s onboarding?
The sign-up form itself looks ordinary because users expect it to work. Value becomes visible after they return to the application and open the user menu, discovering self-service profile, password and 2FA controls. We had known that moment for years, so the product work was continually removing time and friction before it.
How did you evaluate tiny developer-experience changes?
At scale, completion data can show whether setup improved. Early on, we watched people's faces while they tried the product, noted the point of confusion, changed it and tested again with somebody new. Removing one prop may look trivial but can require months if the system underneath must change.
Why is product context hard to transfer as the team grows?
The early team heat-seeked from one improvement to the next without documenting why every convention existed. A new hire may copy an established pattern that insiders already know is sandpaper waiting to be fixed. We are writing the principles down, but transmitting which patterns are ideals and which are compromises remains difficult.
How does founder obsession shape the product?
I still go deep on cookies, authentication and details such as whether a component needs a particular prop, sometimes after helping a customer late at night. That expertise drives Clerk's DX, while other leaders need room to develop equally strong ownership of their areas. Outstanding quality requires organizational support for those obsessions, not necessarily the founder doing everything.
Why is developer marketing unusually hard to hire for?
The best work needs both communication skill and a developer's lens: somebody who understands the obscure browser change, framework constraint or setup failure behind a story. That context is difficult to learn from a generic list of marketing best practices, which is why a technically experienced marketer can be unusually valuable.
Why park a Clerk van outside Y Combinator?
YC produces many companies that all need authentication, but Clerk was not a YC company and lacked its internal distribution. With nearby advertising space unavailable, we wrapped a van, arrived early enough to park outside and made sure founders at least recognized the name. Because it moves, the same local campaign can appear at GitHub, Cloudflare and community events.