Your channel mix is over-indexed on the loudest platform, not the one with the best intent.

Why last-click attribution starves your best-intent channel, and the 90-day test that exposes it.

If you're at $1M to $10M, your paid social costs are climbing, your TikTok output is growing, and the plan for next quarter is probably to spend more on what you're already doing. Before you scale that spend, audit your channel mix, because there's a good chance you're over-indexed on the loudest platform and under-invested in the one with the strongest purchase intent.

The two platforms do different jobs. TikTok is top-of-funnel entertainment. It builds awareness and occasionally goes big, but it's expensive, fragile, and relentless, you're always making the next thing, and the cost per result tends to climb as you scale. Pinterest does something structurally different: it catches demand that already exists. People search it with intent, by occasion and need, and the content has a half-life measured in months because it's search-indexed rather than feed-dependent. For occasion-driven, visual fashion, that intent is exactly the demand you most want to capture.

So why does Pinterest stay a rounding error in most growth-brand budgets? Usually not performance. Attribution. If you judge channels on a last-click model, the platform that earns the final click gets the credit and the budget, while the channels that assisted, that planted the intent and warmed the buyer, look underpowered and get starved. Pinterest is frequently an assist channel: it influences the consideration and then the customer converts via a branded search or a retargeting touch that takes the last-click credit. Measure only last-click and you will systematically underfund the channel that's quietly doing upstream work.

The correction is not to abandon TikTok or to go all-in on Pinterest on a hunch. It's to stop flying blind on the mix and let real data decide. Reallocate a fixed, contained slice of test budget to Pinterest for 90 days. Build proper keyword-rich pins linked to the right product pages, organized by how your customer actually searches. Then measure assisted conversions and view-through influence, not just last-click revenue, and compare the fully-loaded cost per acquired customer against your TikTok and paid-social benchmarks. Ninety days is enough to see whether Pinterest's intent and durability are doing more than your last-click dashboard has been giving them credit for.

This quarter, run a 90-day Pinterest reallocation test with a fixed budget and judge it on assisted-conversion and fully-loaded CAC data, not on the last-click numbers that have been hiding its contribution.