Milin
Desai

Desai made a decision early on that sounds like it came from someone who'd never taken a pricing course: charge approximately one-third of what competitors charge. Not as a temporary tactic. As a permanent philosophy. The competitors are watching their margins; Desai is watching his developer love-to-customer conversion rate.

The logic is upstream of pricing. If you're building for developers - the people who actually choose and champion tools in their organizations - then friction at the point of adoption is the real enemy. A tool that's obviously worth paying for, priced in a way that doesn't require a procurement committee, spreads through organizations at a fundamentally different rate than an enterprise solution that needs a VP signature. Sentry's paid customers often started as developers who just tried it.

When investors push on margins, Desai pushes back on the framing. "It's packaging, not pricing," he's said - the real question isn't what to charge but which features belong at which tier. Get that right, and price becomes a downstream variable. Get it wrong, and you can charge anything you want and still lose.

He's also turned down enterprise deals. Not because they were too small - because they pointed in the wrong direction. Companies that wanted Sentry to compromise its developer-first posture in exchange for a larger contract didn't get the contract. This is the kind of decision that reads as obvious in retrospect and terrifying in the moment. At a company with VC backing and growth expectations, saying no to revenue requires conviction in the model that most executives don't actually have.

Sentry was open source before Desai arrived. He decided it would stay that way - not out of ideology alone, but because he understood what open source actually buys you. "The reasoning was: if we deliver value and do it well, people will want to pay, one way or the other." This is a cleaner version of the freemium thesis than most companies execute, because it ties the open version to the commercial one through genuine value rather than artificial limits.

What Desai doesn't accept is the asymmetric version of this arrangement: large companies consuming open-source infrastructure without contributing financially back to maintainers. He's been vocal about this publicly. The critique isn't abstract - it's about the sustainability of the ecosystem that tools like Sentry depend on, and the precedent set when consumption is divorced from contribution.

When every software company is describing itself as an AI company, it helps to have a number. Desai has one: 95% root cause accuracy, achieved by feeding Sentry's production context - the actual stack traces, breadcrumbs, and event data - into large language models. In error monitoring, root cause identification is the hard part. The stack trace tells you where the crash happened; the root cause tells you why it happened and what to fix. Getting that right 95% of the time changes the economics of debugging.

Desai's view on AI adoption is characteristically direct: "The question isn't whether to adopt AI but how to use it responsibly and effectively to transform your business." He's positioned Sentry as an AI-native platform, not an AI-augmented legacy tool - a distinction that matters for how the company thinks about roadmap priorities versus retrofits.

His appearances at AWS re:Invent 2025 and on the First Round Review podcast in early 2026 have centered on exactly this: what "becoming AI-native" means for a developer tools company that was already doing sophisticated event processing before LLMs entered the picture.