On July 30, StackAdapt announced that its self-serve platform now plugs directly into Affinity Solutions' transaction ledger: 150 million cards, 100 million consumers, 86 billion transactions, north of $4 trillion in spend, roughly a quarter of the U.S. adult population. Advertisers can now watch a campaign run and see, inside the same dashboard, how much card spend followed it. That same week, Forrester published a forecast nobody at Cannes was quoting on stage: confidence among marketing leaders in their ability to measure their own impact is set to fall in 2026, for the first time in years. Only 72% of B2C marketing leaders say they can demonstrate business outcomes with confidence. Just 30% feel sure they can measure aggregate ROI across the channels they actually run, and the average marketer is now running fifteen of them.

Why does $4 trillion in data still not equal a causal answer?

Because a bigger denominator doesn't answer a different question. StackAdapt's tool measures what happened after exposure, total attributed visitor spend, transaction counts, origin markets. That is a correlation, dressed in the confidence of scale. It tells you a card was swiped near a campaign, not that the campaign is why. The tool itself launched aimed at travel and tourism marketers first, where the story is easiest to tell: someone sees an ad, books a flight, the geography lines up. Try the same logic on a CPG brand running fifteen channels at once and the story stops holding still. This is the exact gap the industry has been circling for two years, MMM came back into fashion specifically because it forces a counterfactual (what would have happened without the spend) instead of a coincidence (what happened alongside it). A dataset covering a quarter of American adults is genuinely useful. It is not, on its own, a fix for the thing marketers say they're actually missing.

Why does your brain buy the story before it checks the math?

Because narrative closure is cheaper than uncertainty, and your brain always takes the cheaper option first. This is the well-documented narrative fallacy: humans compulsively convert sequences of facts into cause-and-effect stories, even, especially, when the underlying process is closer to noise. Researchers studying causal reasoning in both humans and language models find the same failure mode: temporal precedence (this happened, then that happened) gets silently upgraded into causal claims, because a confident story resolves the discomfort of not knowing faster than a rigorous test does. A CMO staring at fifteen unmeasured channels isn't choosing between certainty and uncertainty. They're choosing between a dashboard that hands them a clean story and a holdout test that takes six weeks and might say the channel did nothing. Every attribution vendor since cookies started dying has been selling the same product underneath the product: relief.

What does the wrong story cost, in budget terms?

It costs you the channels that can't tell a fast story, which are usually brand and creative, not performance. IAB's 2026 outlook has advertisers raising spend 9.5% for the year, and even with 90% of buyers naming tariffs as a live threat, the share actually cutting budget in response dropped from 45% to 30% year over year. Where is the money moving instead? Toward "channels with more robust measurement capabilities", the industry's own phrase for channels that produce a story fast. Social is up 14.6%, connected TV 13.8%, commerce media 12.1%: the three categories best wired for exposure-to-purchase narratives, StackAdapt's new travel use case among them. Upper-funnel and brand spend, the harder-to-attribute half of the budget, gets starved not because it works less, but because it explains itself more slowly. Run the CPM math on that shift and you're not funding what performs, you're funding what narrates.

Attribution tools don't remove uncertainty from a marketing budget. They relocate it, from a CMO's gut feeling to a dashboard number, and a dashboard number is much harder to argue with in a board meeting, whether or not it's true.

What this means for your next budget cycle

Before you buy the next platform promising to finally "prove ROI," ask it the one question its sales deck will dodge: what's the counterfactual, and who ran without the campaign? A transaction dataset without a holdout group is a bigger mirror, not a better lens. Put a small, explicit line in next quarter's plan for incrementality testing on your two least-measurable channels, the ones currently losing budget for being slow to explain themselves, and let the geo-holdout or matched-market test argue for or against them on its own terms. The brands that win the next two years of budget fights won't be the ones with the biggest dataset. They'll be the ones who can say, out loud, which numbers they haven't tried to force into a story yet.