Creators are now a must buy for nearly half of ad buyers. The IAB, with Advertiser Perceptions, found 48% of ad spenders consider creators a must buy, behind only social media and paid search, and Disney says more than 40% of the assets in any of its media campaigns are now creator-led, up from probably under 10 percent five years ago.
And yet the IAB’s own measurement review says budgets “continue to grow based on directional confidence rather than mathematically provable incrementality.” That is the thing worth sitting with. The channel with the least proof is taking the fastest-growing share of the budget. The answer is not to pull back. It is to measure creators the way you measure any channel with real money in it.
Creators became a line item, not a test
The spend tells the story. The IAB projected US creator economy ad spend at $37 billion in 2025, a 26% increase and about four times faster than total media, and it has grown from $13.9 billion in 2021 to a projected $44 billion in 2026. That is a line item now, not an experiment, and brands are signing long-term creator deals instead of one-off posts. Chili’s said creator content “performed so much better than the spot that we spent so much time creating.”
But notice what brands say they are buying. The top stated creator campaign goals are brand awareness at 43% and reaching new audiences at 41%, with driving online sales or conversions fourth at 32%. Those are upper-funnel goals, which means creators are mostly demand generation, not demand capture, and should be judged on whether they created demand, not on whether they were present when it converted.
That mismatch between what brands buy creators for and how they grade them is the whole problem in one sentence. The budget is approved on awareness and new audiences, which are demand-creation goals, and then the results are read off promo codes and last-click conversions, which are demand-capture metrics. You end up judging an upper-funnel investment with a lower-funnel ruler, and the ruler will almost always say the creator did less than it did, or more, depending on who else touched the sale. Either way the number is answering a different question than the one the budget was approved to answer.
Every creator metric is graded by someone who benefits from it
Here is why the measurement has not grown up with the budget. Almost every creator metric is reported by a party with a stake in the answer. Views and engagement are counted by the platform that sold the placement, and the IAB’s January 2026 review is blunt that engagement “is not a business outcome”, it is an input signal. Promo codes and affiliate links credit the last touch, which the same review says “disproportionately credit last-touch behavior.” Earned media value does not measure sales, lift or incrementality at all. And TikTok Shop GMV is a platform reporting its own sales.
Last-touch credit has the same flaw as retargeting: it takes the bow for a decision made earlier, the mirror effect where the channel nearest the sale claims it. Push it far enough and people engineer being last. In the clearest example, creators allege the Honey browser extension replaced their affiliate cookies at checkout, a class action filed in December 2024; PayPal says it follows last-click attribution. When the commission goes to whoever touched the cart last, the incentive is to touch it last, which is the whole problem with letting the seller keep the scorecard.
None of this means the engagement is fake or the GMV did not happen. The sales are real. The question is whether they are incremental, and no metric that is counted by the party being paid can answer it. That is not a creator problem. It is a measurement problem that creators inherited when they became a budget line instead of a test.
The holdout that should scare you, and the one that should reassure you
The only way to settle it is to turn the channel off and watch. In one vendor-published case study with an anonymous brand, Haus reported that a global DTC brand switched off all its US affiliate loyalty partners for seven weeks and saw no change in revenue against comparable regions, even though affiliate’s share in Google Analytics dropped sharply. The brand then significantly lowered its US investment in that channel. The sales it was paying for were already happening.
To be fair, the test cuts the other way too. The IAB argues that “creators don’t always close the sale, but they prime the system that closes the sale,” so last click can undervalue them. And in a 2024 Journal of Marketing study using sales data and field experiments, smaller influencers delivered significantly higher returns from direct sales campaigns. The point is not that creators do not work. It is that you cannot tell under-credited from non-incremental without a holdout you control.
Pay for new customers, not for the sale
Treat creators like any other channel with budget behind it. Four moves.
First, judge them on new-to-brand customers and on what those customers do next, not on GMV, because discount codes from creators tend to bring in discount-shaped cohorts that retain worse. Second, hold out a region, an audience or a time window before you scale, and read the whole business, not the creator dashboard. Third, add a post-purchase “how did you hear about us” question, which catches the influence last click misses. Fourth, separate creator content as an asset, reused in your own paid media the way Chili’s did, from creator posts as media; those are two different things and need two different scorecards.
I have run this discipline at scale. At a large retailer where I owned a $36M annual paid media budget, we did not take any channel’s own report at face value. We read platform numbers next to a post-purchase survey that asked buyers how they heard about us, and next to incrementality tests. The triangulation is what kept a confident dashboard from deciding the budget.
Price discipline on the most discount-driven shelf there is
Creator commerce leans hard on codes, and the scale is real. Shapermint is on track for about $10 million in TikTok Shop sales this year with more than 5,000 active affiliate creators, 60% of its top performers over 35, targeting women 35 to 65. That is a channel, but the article makes no claim about incrementality, and neither should you.
There is another path. Litter-Robot, whose unit sells for $699, is running TikTok Shop with firm pricing and niche affiliates, moving buyers into its own app, and Whisker expects a halo on DTC and Amazon, though that halo is an expectation, not a result yet. Full price with the right creators is possible, and it protects more than margin. A creator program built on codes is training your audience to wait for the code, the same trap as the luxury playbook for a $99 dress. I have made that reallocation myself. At a subscription wellness brand I ran as a fractional head of growth, I exited paid influencer spend that could not show new customers and moved the money into affiliate and podcast host-read. Affiliate grew 240%.
Live events are the honest version of creator measurement
The cleanest creator measurement is the kind the platform does not touch. Dude Perfect put more than 200,000 fans through stadium gates in 2025 and judges its tours on ticket and merch sales, where a sponsor can tally the turnout, sales and sign-ups before the night is over. Turnout, merch and sign-ups are counted by the person who paid, not by the party selling the reach, which is exactly where live events earn their place in a measurement stack.
Creators earned the budget. Now they need the same test every other channel has to pass. If you turned your creator program off in one region for six weeks, what do you expect would happen to new customers?