The toolkit didn't survive the speed: roadmaps, OKRs and two-week sprints
The cult of productivity has reached a new level of madness.
The norm has shifted to one hour
Barely a year ago, "I'll have it by tomorrow morning" was an absolutely normal answer. What's more, it counted as fast. Today that same phrase reads as "get lost, I'm off to have tea". The norm has shifted to one hour, and one hour is already stretching it, because around the fortieth minute a "how's it going?" is going to land on you.
The most disgusting part, and I really don't feel like admitting it: I have become the source of that same madness myself. I catch myself pinging people after an hour and getting quite sincerely irritated that they haven't answered me yet. With my head I understand it's inadequate, and I still demand it. The speed of the market dictates the new conditions, so that stupidity has become an internal standard.
What broke is the tools, not the people
The real problem is that at that speed the instruments everybody got used to over the years have come apart. And half of them, to be honest, never even got round to learning how to work with them: quarterly plans, roadmaps, demos, retros, two-week sprints, the sacred "let's do grooming next week". All of that is methodology from a world that no longer exists. While you approve the quarterly plan, the market has time to change its shoes three times. Half a quarter has gone by for me and in some directions I have already corrected the OKRs.
That is the honest verdict on quarterly planning: not that agile methodology was wrong, but that its unit of time was calibrated for a market that moved slower than this one. A product roadmap that survives a quarter is now a description of a company that stopped looking outside.
How I reorganised the work
- A hard MVP for every task. Not "I'll do it well and show it", but "I'll do the absolute minimum and show it today". Eighty per cent of the polishing goes in the bin anyway after the first contact with reality, you just find out two weeks later.
- The shortest possible iterations. In half the cases the iteration is one day.
- A definition of done for everything. Otherwise the task spreads like dough and eats exactly the time you gave it, and you gave it by eye and very generously.
- Batching by context. The jumps between analytics, hiring and calls cost more than the tasks themselves: more goes into the switch than into the work.
- All the stupid routine handed to the neural networks. These devil machines cannot be allowed to think for me, but unloading my hands is compulsory. That is exactly what automation is for.
Sprint planning survives this list only in the sense that something still has to be decided first. Most prioritization frameworks don't: they were designed to rank a backlog that stays still long enough to be ranked.
Where it shows best: hypotheses
Around ten a week go through my RnD. Before, each one waited for a full justification before it was even picked up. Now the hypothesis goes into battle with a minimal cut of data, and either we quickly see that it's dead, or we tune it on the move. That works only because research is counted in hours and not in weeks.
Conclusion
For all the smart alecks in the comments, in advance: everything that's happening is not a victory over the environment and not a way to show that you're all doing it wrong while I stand here dressed in white. It's simply a way to survive in this environment, not go mad and stay competitive. Nobody is going to rewind the speed, and the world is not going to regress. Strategic planning didn't die; the calendar it was written on did.
