Walk one use case yourself, end to end

Here is the exercise I ask every IT leader to do before we talk about tools, vendors or operating models. Pick one small use case, something a product team wants and could safely test. Then stay with it from beginning to end, personally, however senior you are.

That means the requirements conversation, the Jira project, the development environment, the repository and the pipeline, testing, user acceptance, the architecture review board, the change approval, the infrastructure request and the production deployment. Do not delegate the walk. Sit in the meetings. File the tickets yourself. Write down every step and every wait, with dates.

What you end up holding

You will finish with a document nobody in your company has ever produced: the true path from an accepted idea to a live user, with the calendar attached. The pain points will not need a consultant to find them. They will be the places where you, with your title, still waited.

That document answers the question most leadership teams keep circling: why the company has stopped innovating, or has started to struggle to deliver. The answer is almost never the people.

Three numbers

Then measure three numbers on that one use case:

• The days to first live exposure, even at half a percent of traffic.
• The number of manual approvals required.
• The steps a developer had to take to get a test environment.

Do the hard, unglamorous work of removing one approval, one step and one day. You will feel the system move.

Speed and control

Stability matters. Compliance matters. Security matters more than ever. But speed matters too, and not for vanity. It matters because learning early is the only reliable way to make better bets. The goal is discipline without the friction, and when IT becomes a platform for that, innovation stops paying the tax.

I wrote The Seven Frictions to give that walk a vocabulary. If you do it and want a second pair of eyes on what you found, send it to me.