One of advertising's most repeated rules says growth comes from reaching the most people you can. This is a small, honest model of whether that holds. Each test runs two campaigns on the same budget that differ in exactly one way — so whatever wins, you know why.
Reaches as many people as it can — but each person sees the ad only once (frequency capped at 1).
Same audience, same budget — concentrates on fewer people, each seeing the ad several times.
Both buy the same inventory, with the same money and the same objective. The only difference is reach versus frequency — many people once, or fewer people repeatedly.
drag memorability, then flip what you measure — on total people made aware, how memorable the ad is decides the winner
Per person reached, more exposures mean more chances to actually be watched, and memory builds with each watched view. This is the number brand-lift studies report, and frequency wins it at any level of memorability. Now switch to “Total people made aware.”
These figures are illustrative. On lift per person reached, frequency always wins: more views can’t lower recall. On total people made aware, the winner rests on two assumptions — how often one exposure is actually watched, and how many watched views an ad needs to stick. For a forgettable or average ad, frequency wins. As the ad gets more memorable the gap closes, it’s level at about 80%, and only a very distinctive ad lets reach edge ahead — narrowly.
A fixed budget buys a fixed pool of impressions, and every impression is either a new person (reach) or another look from someone already reached (frequency). You are only ever dividing the same pool — the test is which way of dividing it builds more awareness.
At a 24% baseline this is a challenger — most people have never heard of it, so there's plenty of awareness to win. Awareness is accrued over many flights, not extracted in a single campaign — which is why one study's lift understates real brand-building.
Two campaigns on Meta, same three 15-second videos, same budget and flight, both optimised for 3-second video views. A brand-lift study on each against one shared control of 500 unexposed, 500 exposed per campaign. For the frequency test, cap A at a frequency of one and leave B uncapped — the only difference is how often each person sees it. For the targeting test, cap both at one and switch A to optimise for cheapest reach instead — now the only difference is who sees it. If reach wins the way the rule predicts, cleanly, past the margin 500 per cell can resolve, you have evidence for it on your own brand. If it doesn't, you have something better than the rule: your own result.