booked 5 vacation days for launch week - what actually needs me awake and what is theatre

Took the week of. Wrong choice. The work that mattered was the week before - writing the emails, the FAQ, the fifteen replies I'd end up sending anyway, the 'sorry, that's broken, here's the workaround' template. On launch day itself I did ninety minutes of useful work and eleven hours of refreshing. Second time round I took the Thursday and Friday before and only the launch day itself off. Better in every measurable way.

Rough split from two launches: prep 70%, launch day 10%, week after 20%. The week after is the one people under-budget. That's when the trial-to-paid conversations happen and when the second wave of 'I signed up Tuesday and got stuck' lands. If you can only move one day, move it seven days later.

Mild disagreement with the emerging consensus. Not because the work is there on the day, but because being reachable inside ten minutes on day one is worth real money when you have single-digit customers. I answered a pre-sales question in four minutes on launch day and the person said afterwards that's why they said yes. You cannot do that from a standup.

Freeze the code Friday. Ship nothing the day before. The single biggest source of launch-week chaos is the confident small change made at 10pm the night before.

Did this in March so it's fresh. What I actually used the days for, ranked by how useful it turned out. One, sat and watched three people I know use the signup flow on a call, which found two dead ends in twenty minutes. Two, wrote seven canned replies in the help widget. Three, went through my own product as a brand new user with a fresh address and found the password reset link pointed at localhost, which would have been a very bad Tuesday. Four, went for a long walk because I had nothing left to do.

The walk was genuinely part of it. Around 4pm on day one my brain wanted to add a feature and the walk is what stopped me.