Why everyone is coming back to the mid-funnel (and how to sell it internally)

For years, the ad market ran on a fairly clean split. TV and out-of-home built awareness. Google, Meta and Amazon captured intent and drove conversions, with metrics (clicks, ROAS, CPA) that came from tools grading their own homework. Those metrics at least existed, and they were easy to get.

In between (and I'm barely exaggerating) sat a no man's land that marketers gradually walked away from. It was the so-called "consideration" stage, once also known as the "mid-funnel." It mostly got pushed to the back burner. Not because it was useless, but because it was hard to put a value on.

That setup is now reaching a turning point. Even with budgets under strain, the tradeoffs no longer come down to a wholesale shift to the bottom of the funnel. A large share of marketers expect budget cuts in 2026, but the split between brand and performance spending is broadly holding steady, according to research firm Warc. Some mid-funnel channels, such as video and creator marketing, are still seen as growth areas.

For many advertisers, 2026 will mark the mid-funnel's comeback. Two forces are driving it: the structural decline of effectiveness measurement and, more importantly, the spread of discovery interfaces dominated by conversational AI. Yes, I mean ChatGPT, Gemini and the rest.

Attribution as the industry sold it is getting harder to defend. GDPR, consent, the limits Apple's App Tracking Transparency (ATT) puts on iOS tracking, disappearing identifiers, fragmented journeys: the ecosystem has become much harder to read.

The fallout shows up in how marketers steer their spend. More advertisers are going back to approaches such as marketing mix modeling (MMM). It relies on correlations between spend levels and overall sales rather than on tracking each user's path. Attribution, especially last-click, is falling out of favor.

MMM accepts a simple reality. You can still optimize, but expecting a complete, deterministic read on every conversion has become wishful thinking. Ballpark measurement is acceptable again, as long as it keeps the business on track.

That gives the mid-funnel new value, because it feeds the bigger picture: brand preference, future intent and halo effects across channels. It matters even more now that many brands (you've probably all read the Nike "case study") are pulling back a little a lot from the "all-performance" model.

E-commerce is the clearest case. Brands that put most of their spend into Google Ads or Meta are now under pressure. They never built a value proposition or brand preference beyond retargeting and promotions. You can probably think of a few examples!

The point isn't that acquisition no longer works. It's that it has become more expensive, less stable and more dependent on outside factors. That leaves brands that neglected the mid-funnel badly exposed.

The spread of ChatGPT, Gemini and other AI assistants raises the stakes, because it changes how discovery works. People no longer just browse from one site to the next. They ask for an answer, a comparison, a shortlist. The journey gets less linear and less "clickable." It also depends heavily on a third party, the assistant, which brands now have to convince to consider them.

Performance now depends less on how widely ads run and more on whether brands can give AI systems reliable, concrete and useful signals. The goal is no longer just to maximize media weight. It's to get recommended, and to come across as a credible option when the user asks: "What's best for my use case X?"

Power is shifting to the consumer. Every use becomes content. Every review or experience becomes a signal. Every opinion feeds the models and weighs on purchase intent. That's also why the creator economy is booming, as we cover in another report.

Brands that break through won't just be the ones buying the most media. They'll be the ones whose products spark conversation on their own, along with authentic feedback and real proof of use. All of that shapes how AI systems rank brands, and whether they surface one... or make it invisible.

The mid-funnel comeback isn't just a debate about formats. It's also an organizational problem. On the advertiser side, a strict split between brand and performance teams gets in the way, because effectiveness now depends on longer, more consistent sequences.

On the agency side, the channel-by-channel approach (social here, search there, creators somewhere else) is hitting its limits as effects combine across journeys. Agencies need to rethink how they approach performance from the ground up. That means building strategies around the signals people produce, and building ecosystems where reach is no longer the only yardstick.

Operationally, the mid-funnel has to be treated as a value chain in its own right: reassurance content, proof of use, verified reviews, presence wherever comparisons get made. Most of all, it calls for hybrid measurement that combines macro models (MMM), incrementality tests and quality metrics, rather than relying solely on last click (hallelujah).

If this trend holds, the mid-funnel could once again become the main place where brands set themselves apart. The reason is simple. It's where preference gets built before intent shows up.

Platforms remain essential for conversion, but they can no longer carry growth on their own. In that world, consideration becomes a strategic investment again. And as conversational AI steers a growing share of discovery, producing credible signals (usage, feedback, recommendations) becomes a performance lever in its own right.

So in 2026, the question won't just be "How much am I paying for media reach?" It will be "What am I putting into my media to deserve a recommendation?"