There are 19,000 zip codes in India and 5,764 municipalities in Brazil. Fernando Fernandez knows the exact count because he wants an influencer in every single one of them. "In some, I want 100," he told PR Week last year.
Since taking over as Unilever's CEO in 2025, he's grown the company's creator roster from around 10,000 to roughly 300,000, with social spend now eating half of Unilever's media budget. Total marketing spend has climbed from 13% of revenue to 16%, which on a company doing €60.8 billion a year works out to a war chest north of €9 billion. "The lifespan of a video is four days," Fernandez says. "You have to create a lot of stuff."
This is a company betting billions that consumer trust in a micro-influencer on TikTok can do the job that eighty years of brand advertising used to do. I get it, media consumption patterns change, big swings happen. What I can't get past is the timing, and the institutional denial of what we know actually builds brands and drives growth.
About a year after the Unilever news, Mark Ritson and Byron Sharp stood on a stage at Cannes and did something neither has done in a decade: agreed with each other, on camera, about what actually builds a brand. Sharp is the Ehrenberg-Bass guy, the "mental availability" guy, the man Ritson has spent years needling in his columns. They still disagree about plenty. But on this they were dead level: brands grow through broad reach and distinctive assets used with what Ritson called ruthless consistency. "Codify the shit out of everything," he said. A logo, a color, a sound, repeated so relentlessly that a half-distracted person, per his own example, talking to the dog, still clocks it. Sharp backed him with the ugly little stat that keeps those of us in marketing grounded: after watching ten ads, only 16% of people can even link the ad back to the right brand. That's the bar. Under ideal conditions, with one message repeated on purpose, four out of five people still get it wrong.
The industry cheered this détente as it should have. Two guys who've spent careers arguing finally settled on the same floor. And then, in pretty much the same news cycle, Unilever doubled down on their commitment to tactics that have yet to prove they build brands and demand at any reasonable, sustainable scale.
Consistency is the mechanism, not a nice-to-have, per years of quantifiable research and the two men who just agreed on stage that it is. So how does a brand stay consistent across 300,000 creators, each with a different face, a different editing style, a different voice, cycling through content with a four-day shelf life because Fernandez himself says the model requires constant renewal? Under that model, a brand looks like 300,000 different things, on purpose, at volume, forever, the precise flip of what the room just agreed matters, applauded by a lot of the same people who were sitting in it.
Their trumpeted value of trust is even worse than the consistency problem.
Clutch surveyed consumers on the topic of trust just last year. 53% said they trust a product recommendation less once they find out the influencer got paid for it. Nearly half hadn't bought anything off an influencer's word in the past year.
This "sponsored blindness," the influencer-era cousin of banner blindness, has one estimate putting the share of users who reflexively scroll past anything tagged "paid partnership" at 60%. The real weak point mechanism behind this shift to embracing the trustworthiness of influencers is that those who take too many deals that come their way burn the exact credibility that made brands want them in the first place. Leading to a scaling problem baked into its own success condition. The plan for more creators, more content, more velocity is also exactly what accelerates the trust decay. You can't out-volume a phenomenon that gets worse with volume.
One more number, because I can't not include it. Last year, at an event Unilever itself sponsors, CreativeX presented an analysis of 1.6 million creator ads representing close to $2 billion in spend. Their finding: roughly half of it is wasted. Same as it ever was with traditional channels but without the consistency and memory building mechanisms. Worth sitting with as well is that Ehrenberg-Bass, the institute behind the exact research now getting quietly sidestepped, is supported in part by Unilever's own headquarters. They're paying for the research and betting against its central findings in the same fiscal year. Make it make sense.
None of this means influencers don't work. They're genuinely good at what some advertising is bad at: demonstration, discovery, getting a product into a conversation it wasn't otherwise going to enter. Nobody serious is arguing the other side of that.
The argument is narrower and, I'd say, more obvious than it should have to be: don't ask a channel that structurally cannot deliver consistency, and is measurably losing the trust it's being purchased for, to replace the two things every serious study on brand growth says are non-negotiable. The best Unilever can offer up so far is higher view-through rates, not increased penetration, revenue, or consideration.
If Unilever has evidence that 300,000 fragmented voices can build memory structures the way seventy years of consistent, distinctive, boringly repeated brand assets did, they should publish it. Considering their aggressive messaging of their budget and focus shift this past year, they most likely would share it if they had it. Cannes would show up for that talk too.
Until then, Unilever is placing one of the largest bets ever against the one idea the industry just spent a stage, and decades, agreeing on, running at full speed in public with the same confidence they had with their purpose-led rebrand of roughly 400 brands. Which worked up on paper until it didn't in fact and investors made clear they cared more about results than headlines.
*** Views expressed on Push Pull Pivot are personal and shared in an individual capacity. They do not represent those of any current or former employer ***