Reuse beats creation: building a library people actually use

Part of a single argument: the engagement gap is a workload problem.

Teams that run a lot of sessions almost never have a working shared library. Everyone has a folder, the folders are effectively private, and the same explanation of the same thing gets rebuilt several times a year by people sitting near each other.

Most organisations have tried to fix this at least once. The attempts fail in consistent ways, and the failures are informative.

Why libraries fail, precisely

Contributing costs more than not contributing. If putting something in the library is a separate act performed after the real work is done, it competes with going home. It will be done enthusiastically for about three weeks.

Finding is unreliable. Not impossible, unreliable, which is worse. An unreliable search has an uncertain payoff, and under deadline pressure people choose the certain two hours of rebuilding over an uncertain twenty minutes of looking.

Results cannot be judged. Even a successful search returns a list of filenames, which tell you nothing about whether the thing is any good or current. The evaluation cost gets added to the search cost, and rebuilding wins again.

Notice that none of these is a discipline problem, and none of them is fixed by a policy asking people to file things properly.

Make contribution a side effect

The only libraries that stay current are the ones where doing the work is the act of contributing. If saving your session puts it somewhere findable by default, the library fills itself and nobody has to be reminded.

This is worth designing for even at some cost elsewhere. A library that captures eighty per cent of work automatically beats one that captures a hundred per cent when people remember, because in practice the second captures about fifteen.

Make finding beat rebuilding on the clock

The bar is not that search works. The bar is that the whole find-and-evaluate loop is faster than starting over.

That means results need enough context to judge in seconds: who ran it, when, with what size and kind of group, and some signal about whether it went well. A filename does not clear that bar. A filename plus a date and an owner usually does, which is a low bar that most shared drives still fail.

The thing that makes it compound

Every session added to a working library lowers the cost of the next one. That is the opposite of how preparation normally behaves, where each session costs roughly what the last one did.

It also means the return grows fastest for the teams that run the most sessions, which are the teams under the most pressure. Of everything in this blog, this is the intervention where the benefit appreciates rather than staying flat, which is why it is worth doing first even though it is not the most broken thing you have.

A smaller version that works

If a proper library is not achievable this quarter, there is a cheap approximation: one shared folder, one naming convention, and a rule that the person who runs a session puts the file there before they close their laptop.

It is worse than the real thing in every respect except the one that matters, which is that it can be started on Monday. Teams that do this typically recover a meaningful fraction of the duplication cost within a couple of months, which is usually enough to fund the better version.