Skip to main content
Predicts, for each supporter on a mailing list, how likely they are to respond to a direct mail appeal — so you can mail fewer packs without losing income. What it predicts — The probability that a mailing leads to a donation within a set number of days of the mail drop. Who it scores — Everyone on the mailing universe: the full list you could mail. The decision it drives — Rank the mailing list and cut low-probability names, so you post fewer packs without losing meaningful income. Worked example — You plan a 100,000-piece appeal. The model ranks the list; the bottom 30,000 contribute almost no expected response. You mail the top 70,000, save the print and postage on the rest, and hold response roughly flat. What good looks like — Cost per response falls and net income holds or rises.
Pairs well with Expected amount — rank on expected revenue (likelihood × value), not likelihood alone.

All models

Back to the full catalogue.

Create & train a model

Set one up on your own data.