The Kept Decision

The visible mark of a well-run scale-up is the decision it deliberately keeps

Ask any company that raised money in the last year what it has decided not to change.

Call those kept decisions: a human approves every credit decision, or one team owns the data model and that ownership survives two reorganisations. Something is being held constant on purpose while almost everything around it moves, and what is held is a piece of judgement rather than a piece of technology.

The companies that can answer at all are the ones this piece is about. What follows comes from reading them rather than from any measured sample, so none of it is a finding.

Growth converts choices into commitments on its own

The mechanism needs nobody's permission. A choice gets made once for a good reason, something gets built on it, and something else gets built on that. By the time anyone looks again the choice is load-bearing and unwinding it costs more than making it did. It became permanent because the company kept working. No decision was involved at any point.

That happens to most choices a growing company makes, and it runs on a clock nobody set. Months go by. Nothing feels wrong, because nothing is wrong yet, and the choice is quietly becoming the thing everything else assumes.

So permanence arrives either way. The variable is whether anybody chose it.

A kept decision is the same thing with somebody's name on it

A kept decision is that same thing decided in advance and defended afterwards: this one stays, here is why, and here is who gets asked before it changes. A company that can write that down has moved the choosing to the front, where it costs a conversation instead of a migration.

There is a real objection and it deserves an answer. Holding things constant is how companies end up with legacy. Good operators would say the discipline is the opposite one: keep everything reversible and commit to as little as you can. In their experience the expensive mistake has always been committing too early, and they are mostly right. The failure this piece is about runs the other way. A kept decision list is short, and what it holds is a boundary, so a company can keep three of them and still replace everything underneath.

The commitments most likely to erode are the cheap ones to reverse

The instinct is to guard the expensive commitments, the platform and the multi-year contract, because the cost of getting those wrong is easy to see. Sorting decisions by how hard they are to undo is a genuinely good way to decide how much attention to spend on each, at the moment it is being taken.

Erosion works on the other end of that list. A commitment that would take an afternoon to reverse rarely gets escalated, because escalating it would be disproportionate. So it gets reversed in an afternoon, by a reasonable person, under a deadline. Then again. The manual approval becomes an exception for one customer segment, then a batch job that runs overnight, then the way it works. Every step was small and every step was defensible. Nobody decided to remove the commitment, and it is gone.

The property that makes a commitment easy to reverse is the same one that keeps it out of any process organised around cost.

With nobody in the chair, the same process is Decision Debt being taken out

Decision Debt is the accumulated cost of technology decisions made without evidence. It compounds silently, and it is invisible until an event makes it visible.

That is CloudFruition's definition, and the flagship paper it comes from describes the room where the bill finally arrives, on somebody else's schedule. The borrowing happens long before that, at the moment a choice became load-bearing and nobody wrote it down.

The kept decision and Decision Debt are one process seen twice.

What to do with this before the next planning session

Take two minutes and write the decisions your company deliberately keeps. Write the specific commitments: what stays human, what one team owns, what an integration is not allowed to touch, what gets built rather than bought.

Then put the same question to two people who would have to defend those commitments, and compare the lists.

Where the lists disagree, that gap is the thing to write down. A kept decision that lives in one person's head is being remembered, by someone who will be on leave the week the deadline lands.

Of the choices your company made this year, which would you notice becoming permanent, and which would simply have happened?