Every startup is a collection of bets. Some are explicit: we're betting that this market is big enough, that this technology works, that customers will pay for this. Others are implicit and far more dangerous: we're betting that our assumptions about user behavior are correct, that the regulatory environment won't change, that our team can execute fast enough.

The explicit bets are easy to track. You can put them on a slide deck and revisit them quarterly. The implicit bets are the ones that kill you.

Making bets visible

The most useful exercise I've done at any startup is what I call "bet mapping." You sit down with your team and list every assumption your business depends on. Not just the strategic ones — the operational ones too. We're betting that AWS won't raise prices more than 10% this year. We're betting that our lead engineer won't leave. We're betting that the API we depend on will maintain backwards compatibility.

Most of these bets are reasonable. But seeing them all in one place is clarifying. You realize that you're carrying more risk than you thought, and that some of that risk is concentrated in places you haven't been paying attention to.

The hedging instinct

Once you can see your bets, the instinct is to hedge everything. Build abstractions so you're not dependent on any single vendor. Cross-train so no single person is a bottleneck. Design for flexibility so you can pivot if assumptions change.

But hedging has a cost. Every abstraction adds complexity. Every cross-training session takes time from building. Every flexible design is harder to optimize. There's a real tension between managing risk and maintaining velocity.

The art of selective hedging

The skill isn't hedging everything — it's knowing which bets to hedge and which to let ride. Hedge the bets where the downside is catastrophic and the cost of hedging is low. Let ride the bets where the upside of commitment outweighs the option value of flexibility.

This is more art than science. But having the map — seeing all your bets laid out — at least gives you the information you need to make those judgment calls intentionally rather than by default.