Part II — Seeing Organizational Flow · Chapter 31 · 3 min read · First Public Draft

Discovery cost

The Detective Work at the Front of Everything

The weeks each initiative spends establishing the facts before anybody starts, redone from scratch every time.

Reduces frictionCreates valueDecision latency

Before an initiative can start, somebody has to find out what is true. This is the cost I would price first if I could only price one, and it is the one nobody has ever put in a plan.

Nothing can start until somebody has established the facts. Does this already exist somewhere? Who owns the thing we need? Is that a person or a forum, and when does the forum next meet? Where does the data live and what do the fields actually mean? Who has to say yes, and what will they want to see before they do?

None of that is the work. It is finding out what the work will involve — and it gets done fresh, by different people, for every initiative, because nobody wrote it down last time either. The people doing it are usually rather good at it, which is part of the problem: an organization full of competent detectives feels resourceful rather than slow.

A few weeks is unremarkable in a large organization. It appears in no plan, because it does not look like a phase. It looks like getting started.

The version that gives the game away is the proof of concept. A PoC is meant to be the cheap fast thing you do to find out whether an idea has legs — a fortnight, in principle, and the whole point is that being wrong costs almost nothing. So ask how long it took from somebody first saying we should try this to anybody actually trying it. The gap is routinely longer than the PoC itself, and nearly all of it is detective work.

Which produces an effect nobody intends and nobody decides on: it raises the bar for what is worth attempting at all. If finding out costs three weeks before anything begins, small ideas stop clearing it. They are not rejected — nobody rejects them, there is no meeting — they simply never get proposed, because the person who had the idea did the arithmetic in their head and decided not to bother. What survives is the large initiative, which is slower, more expensive, and considerably harder to be wrong about cheaply.

And the facts do not stay found, which is the part that turns a one-off cost into a recurring one.

The version I have seen most often is small and almost polite. Something moves — an endpoint, a report, a field, a shared folder, a responsibility — and the move is announced to the people who were in the room when it was decided. Everybody else finds out by trying to use the old thing and discovering it is not there.

Nobody hid anything. There was probably even a message, in a channel, in March. But a change communicated once, to whoever was present, is not the same as a change anybody downstream can find, and the second group is always larger than the first.

The other version is quieter still: the thing is exactly where it was, and the way in has gone. A credential expires. A permission is tightened during a security review, correctly. The person who granted access last time has moved on and nobody inherited the ability to grant it. The capability has not changed at all — it has simply become unreachable, and from the outside those two look identical.

Both restart the detective work from the beginning, for everyone except the people who already knew. And both are usually discovered by somebody in the middle of something else, which is why they get handled as an interruption rather than recorded as a cost.

Which means the discovery tax is not paid once per initiative. It is paid again every time something moves without leaving a forwarding address.

An organization that cannot afford to try small things ends up making only big bets. That is not a strategy anyone chose. It is what a discovery cost does to a portfolio, quietly, from underneath.

Cultivating this: Opening the Front Door, in Part III.