Les Binet's Cannes 2026 Warning: Budget Beats ROI 9 to 1, and Efficiency Is a Death Spiral

The most counterintuitive talk at Cannes wasn't about AI

While most of Cannes Lions 2026 was busy debating artificial intelligence, Les Binet got on stage and said something far more dangerous to the way most marketers operate. He argued that the entire industry has optimized itself into slower growth, and that the metric everyone worships is the wrong one.

Binet, half of the Binet and Field partnership whose IPA research underpins modern marketing effectiveness, joined TikTok's Adrian Adaramoye for a session pointedly titled Go Big, Or Just Keep Posting?. His answer was unambiguous. Go big.

"We've become obsessed with the small"

Binet's opening diagnosis, as reported by Marketing-Interactive, was blunt: "We've become obsessed with the small."

He went on: "We love clever little tactics and creative little stunts, and we're all trying to do more with less. I think that's killing our industry."

This is the exact trap I describe in Never Always, Never Never. The way most marketers approach their craft today is broken. We've become trapped in a cycle of short-termism, driven by the allure of instant results and the pressure to justify every dollar. But chasing immediate wins is a race to the bottom. Binet just put hard numbers behind the same instinct.

The number that should end the ROI obsession

Here is the line that should rearrange how you build a budget. Binet said: "The main driver of profit is not ROI, it's the budget by a factor of eight or nine to one."

He followed it with the implication most marketers refuse to confront: "Budget is nine times more important than ROI, and that suggests that the most important decision in marketing is actually how much to spend."

Sit with that. We spend most of our time optimizing return on ad spend, tweaking bids, shaving cost per acquisition, A/B testing buttons. Binet's data says the lever that actually moves profit, by an order of magnitude, is the size of the bet itself. You can run a perfectly efficient campaign straight into irrelevance if it's too small to be noticed.

Efficiency up, effectiveness down: the death spiral

Binet's sharpest framing was the contradiction at the center of modern marketing: "The good news is that advertising is becoming more efficient. But the bad news is that advertising is becoming less effective."

Those are not the same thing. Efficiency is doing things cheaply. Effectiveness is growing the business. A tightly targeted, narrowly focused approach increases efficiency a bit, but it reduces effectiveness a lot. Over time, sales and profits fall, budgets come under more pressure, reach shrinks again, and the cycle becomes self-reinforcing. Binet calls this the death spiral, and he summed up the destination in one sentence: "Effectiveness is largely about scale in marketing. Brands that think small eventually become small."

This is the same disease I call out in the book as Maslow's Hammer. When the only tool you have is a spreadsheet, every problem looks like something you can solve by tweaking a cell. If an idea can't be tied to a clean ROAS number inside a 30-day window, we decide it doesn't count. So we starve the strategies that create the most long-term value, simply because they refuse to fit in our attribution framework.

Why narrow targeting quietly fails

The reason efficiency and effectiveness diverge comes back to a rule that runs through everything I write. At any given moment, only about 5% of your audience is in-market. That is the 95/5 rule, and it explains Binet's whole argument.

When you target narrowly for efficiency, you talk only to that 5% who are buying right now. You get a flattering ROAS because those people were going to buy anyway. But you stop building mental availability with the 95% who will be in-market next quarter and next year. You harvest demand without ever creating it. The dashboard looks great, right up until growth stalls and nobody can explain why.

Binet's prescription is the opposite of narrow. "Big brands need big budgets, they need big media plans, and they need big campaigns. If you think small, you'll become small. So think big to stay big."

This is not an argument against performance marketing

Binet was careful here, and so am I. He stressed: "It's not an either or, you need both, because each enhances the other." Brand building and performance are not enemies. Brand makes performance cheaper, because a brand people already recognize earns higher click-through rates, higher conversion rates, and lower acquisition costs. Performance harvests the demand that brand creates. That is the halo effect in action.

The question is the split. Binet's IPA research points to roughly 60% brand building and 40% activation as a typical starting point, though it varies by category. As I note in the book, online-born brands often need to skew even harder toward brand, sometimes closer to 70/30, because digital categories are so crowded that the scarce resource isn't another performance lever. It's memorability.

What to do with this

Binet's final practical note was about persuasion inside your own company: "If you want to convince your CFO to release extra budget, then you need a proper financial model." The way out of the death spiral is not a louder argument about brand. It's a credible model that ties marketing investment to profit over a multi-year horizon, the kind that marketing mix modeling and incrementality testing make possible.

If your costs have been climbing while growth has been flattening, you may already be in the spiral. The instinct will be to cut, target tighter, and chase efficiency. Binet's evidence says that instinct is exactly backwards.

This is the core argument of Never Always, Never Never. The brands that win don't optimize themselves into irrelevance. They have the discipline to spend at scale, build demand for the 95% who aren't ready yet, and let the cake bake long enough to matter.

The full framework for balancing brand and performance, escaping short-termism, and building equity AI can't replicate is in the book. [Read Never Always, Never Never](/amazon).