There’s a version of “move fast” that actually costs you time. It’s the version where you skip something because it feels like overhead, and then spend twice as long undoing the consequences.

I did this twice in the last month building Fractional Tools. Both times, I knew I was skipping something. Both times, I told myself it was fine because I was moving fast.

The first one: no staging environment

I was pushing changes directly to production because setting up a proper staging environment felt like yak shaving. I had no users. The friction of a staging branch felt disproportionate to the risk.

Then I shipped a bug that broke the invoice workflow for all three of my early testers at once. It took me four hours to diagnose and fix. It would have taken twenty minutes on staging.

The lesson wasn’t “always have staging.” The lesson was: the safeguard you’re tempted to skip is usually the one protecting the thing you care most about right now.

The second one: skipping input validation on a customer-facing form

I was iterating quickly on a subscriber preferences form. I skipped server-side validation because I had client-side validation and I was the only user.

A week later, I had fifty subscribers. Someone with an unusual email format couldn’t save their preferences. I had no error logging on that path. I spent three hours figuring out what was happening.

Again: not about process for its own sake. About the specific thing I was skipping and why.

The pattern

Both cases had the same structure:

1. I’m moving fast on something I care about
2. There’s a step that feels like overhead
3. I skip it because I tell myself the risk is low
4. The risk materializes at the worst possible time

The skip made sense in isolation. The problem was that I was evaluating “is this step worth it?” at the moment of highest momentum, which is also the moment of lowest judgment.

What I do now

I keep a short list of things I don’t skip regardless of pace. Not a long checklist — a short list of the specific safeguards that have burned me before.

For me right now that’s: staging before production for anything customer-facing, server-side validation on any form that touches subscriber data, and error logging before I call something “shipped.”

Everything else is fair game to cut. But those three are non-negotiable, because I know what happens when I skip them.

The goal is still speed. The question is which safeguards you can actually afford to remove.