Main idea

As long as a note base has one reader and one use, its vocabulary holds up on its own: whoever writes knows what they put behind their words, and nobody else has to guess. The need for a glossary does not appear with the volume of notes, it appears with the number of functions drawing on them — product, marketing, sales, support, leadership, design, engineering — and with the number of deliverables pulled out of them.

It is a threshold, not a continuous growth. Below it, an entry is decorative. Above it, the same words start circulating between different readings: on the marketing side, the customer's words, the internal words, the market categories, the perceived alternatives, the recurring objections, the terms to avoid; on the product side, the signals, the stakes, the decisions, the trade-offs, the constraints, the segments, the usage situations. It is those lists that call for stabilisation, not the general vocabulary of the trade.

Why it matters

This avoids the two timing errors. Opening a glossary too early produces a decorative dictionary, full of correct definitions nobody consults. Opening it too late lets each function install its own reading, and you then have to undo apparent agreements before you can define anything at all.

It also gives the right trigger to watch for: not the size of the base, but the moment a second person or a second purpose comes to draw on it.

Nuances and limits

The threshold can also be crossed alone: a single author producing several deliverables of different kinds from the same notes ends up in the same situation as a team, because the word does not mean the same thing in an article and in a specification.

Conversely, a large but single-purpose team can do without a glossary for a long time, for lack of any divergence to reveal.

Open questions

• Is there an observable sign that the threshold has been crossed, other than the first discussion in which two people discover they were not using the same word the same way?