Recency

Recency is how long it has been since a user last came to the site; it is one of the three dimensions of the classic RFM model, alongside frequency and value.

Of the three RFM dimensions, recency is the strongest predictor of next behavior: someone who came very recently is likely to return, while someone who has not come in a long time is on the edge of churn, even if they were frequent in the past. Recency is therefore usually paired with frequency to form four meaningful groups: recent and frequent (the loyal core), recent and infrequent (a newcomer needing nurture), old and frequent (an at-risk loyalist to reactivate), and old and infrequent (nearly lost). A common trap is setting the recent threshold arbitrarily; the boundary must fit the product's natural cycle — a few days for a news site, a few weeks for a seasonal shop — or everyone looks either recent or old and the split becomes meaningless. Recency is a signal for timing an intervention: the best moment for a reminder is just before the user crosses the churn boundary.