The $2,500 test

TAKEOliver Wakefield-Smith · Digital SignetSources checked July 31, 2026

Cue

How should a small business spend $2,500 on podcast ads?

Split it across two shows, not five: $1,250 each buys 48,077-52,083 impressions per show at the published $24-26 host-read CPM, enough for two to three spots on a mid-tail show. Two shows give you a per-show comparison; five give you noise. Total buy: 96,154-104,167 impressions.

96,154-104,167Total impressions2,500 ÷ 26 and ÷ 24, x 1,000

48,077-52,083Per show at 50/50 split1,250 ÷ 26 and ÷ 24, x 1,000

2-3 spotsPer show, mid-tailAt 15-20k downloads per episode

Why exactly two shows

One show gives you no comparison: a good or bad result could be the show, the channel, or chance. Five shows at $500 each gives you one spot per show, and single spots underperform structurally (frequency). Two shows at $1,250 each is the smallest design that produces a per-show winner while keeping two-plus spots per show. Give each show its own promo code; that is the whole measurement design.

What to pay

Anchor on the published card : $24-26 CPM for host-read 60s, genre band $22-30. On a 15,000-download mid-tail show, one spot at $25 CPM prices at $375 (15,000 ÷ 1,000 x 25). Three spots on each of two such shows: $2,250, inside budget with contingency left. Quotes materially above the band need a stated reason. Where to buy: Libsyn Ads or Gumball; the head-to-head with self-serve alternatives is here.

Expected ranges, and the scaling trigger

At a 1% visit rate, 96,154-104,167 impressions model to roughly 962-1,042 visits; at 2% conversion, roughly 19-21 orders; modeled CAC around $119-130 (2,500 ÷ 21 and 2,500 ÷ 19). Those assumption inputs are yours to set in the Test Modeler, and the output is a model, not a promise. The scaling trigger you pre-commit: if the better show’s measured CAC lands inside your tolerance, take that show plus two like it into the $15k quarter.

Smaller: the $500 test · Bigger: the $15k test · Local variant: local business