The sampling revival.

§I The math has turned.

Sampling, as a channel, went out of fashion inside DTC-era consumer between roughly 2016 and 2022. It was physical, difficult to attribute, and slow to produce a readable number on the dashboard. Performance marketing, by contrast, produced a number you could defend in a Monday meeting. The channel was effectively de-funded, even though the category had been using it for a century to acquire repeat purchase.

In the last thirty-six months, the math has turned. Paid-social CPA has roughly doubled in most of our client categories; repeat rate is harder to buy on a first-touch basis; the algorithm gives a smaller and less predictable window to any given creative. Sampling, which costs roughly the same as it did in 2018, has become comparatively cheap. Our clients who have re-introduced disciplined sampling in the last eighteen months are, on average, seeing a 2.1-to-2.8× improvement in blended CAC across the cohort.

§II Why the growth team will resist.

Sampling refuses to be dashboarded. Its payback is measured in repeat rate at 90 and 180 days, not in clicks. Its attribution is hand-coded or, more usually, unattributed. The growth team, whose performance is measured against weekly dashboards, will correctly notice that sampling makes their dashboard worse, not better. They will resist.

The resolution is not to overrule the growth team. It is to remove sampling from the growth team's budget. Sampling is a brand channel, funded on a brand budget, measured on repeat rate, accountable to the founder. Run it adjacent to the growth team, with its own line, its own metrics, and its own cadence. The growth team's number will remain clean. The brand's number will improve.

Sampling makes the growth dashboard worse and the brand P&L better. If you cannot tell the difference between those two things, your dashboard is running your brand. — Théo Marchetti, retail intelligence

§III How to do it well.

The three disciplines that distinguish successful sampling programs from wasteful ones, in our client base. One: sample to a defined archetype, not to a general audience. Two: sample in the context the product will ultimately be used in, not in an airport or a gift bag. Three: include a hand-addressed note, on paper, with three lines of tonal copy. Not a QR code. Not a discount code. A note.

These three disciplines sound precious. They are the difference between a sampling program that recovers 2.4× CAC and one that recovers 0.6×. We have watched both, repeatedly. The three disciplines together cost perhaps ninety cents per unit sampled. The lift they produce is not comparable to their cost.

Sampling is not a fad revival. It is a return to a channel that never stopped working; the dashboard merely stopped rewarding it. Move the channel off the dashboard, discipline the three habits, fund it from brand not growth, and it becomes, again, the most underpriced acquisition move in the category.