Every retailer already owns the answer to how its product should be sold. It sits with the four or five associates per region who consistently outperform, and it has never been written down in a form anyone else can learn. A behaviour taxonomy is the process of extracting it and making it measurable.

Start from outcomes, not opinions

Identify the top quartile by outcome — close rate, average ticket, attach rate over a period long enough to exclude luck — before listening to a single conversation. Not the associates management admires, not the ones with the longest tenure, and not the ones who present well in a training room. The selection has to come from the point-of-sale record, because the alternative recreates the same folklore you are trying to replace.

Listen for the move, not the words

Across a few hundred conversations from that quartile, the same handful of moves recur, phrased completely differently every time. One associate frames a budget as a range before showing anything; another asks what the piece is for inside the first two minutes; a third schedules the follow-up before the customer stands up. The taxonomy captures the move. It never captures the script, because scripts do not transfer across languages, markets or personalities, and enforcing one makes a good seller worse.

A behaviour is something you can observe happening or not happening. If two careful people can disagree about whether it occurred, it is not written tightly enough to score.

Write the definition until it stops being arguable

  • "Built rapport" is not a behaviour. It cannot be observed, scored or coached.
  • "Asked what the purchase is for, before presenting a product" is a behaviour. It happened or it did not.
  • Each definition needs an inclusion boundary and an exclusion boundary: what counts, and the near-miss that does not.
  • Dispute rate is the quality signal. A definition that generates repeated disputes is badly written, and the fix is the definition rather than the associate.

Quantify the gap so the priority is not a matter of taste

For each behaviour, compare performance rate in the top quartile against the store median, then attach the outcome difference. This turns a list of good practices into a ranked list of what to coach first. It is also what makes the taxonomy defensible internally: the priority order comes from the retailer’s own floor, not from a vendor’s framework.

Keep it small and let it change

Six to ten behaviours per vertical is the working range. Beyond that, a store manager cannot hold them and a daily prompt has nothing to focus on. The set should also be revisited as the assortment, the season and the customer change — a taxonomy that has not been edited in two years has stopped describing the floor it was built from.