Five celebrity brands died. Only one of them did

The list of celebrity brands that failed#

That line is the best thing written about celebrity brands this year, and it deserves better evidence than the list it arrived attached to. A widely shared piece named five brands as shutting down. Checked one by one against primary sources, the list does not hold, and the way it fails is instructive in itself. Follow the trail back to each source and a different animal emerges.

Start with the one that stands up. Más+ by Messi has genuinely been wound down. Mark Anthony Group, the parent, put its chief executive Phil Rosse on the record in May: "Many great milestones were achieved. However, despite our best efforts, we did not achieve all of our objectives." That is a real closure, reported by just-drinks and The Grocer on the same day, with a named executive attached. One out of five.

Now the one that has not happened, and this is where the scent goes cold. Unwell, Alex Cooper's drinks line, was described as discontinued in autumn 2026. Today is the last day of August. What was actually reported in mid-August was a plan: a Halloween-themed release, after which production would cease. As of this morning the Unwell shop is live and selling nineteen products. Writing that as a completed closure turns a stated intention into a fact, which is the single commonest way a business story goes wrong.

Then two that closed, but not in the way the framing implies. The Skkn by Kim site carries the brand's own notice that it wound down and closed on 29 June 2025, and Retail Dive reported Coty selling its twenty per cent stake to Skims two months earlier, with Kim Kardashian describing the move as uniting everything under the Skims brand. That is a consolidation with a relaunch behind it, fourteen months ago, not a market exit.

Flower Beauty is the same shape. Beauty Independent reported in September 2025 that Maesa, the incubator and co-owner, had decided to leave the colour cosmetics category to concentrate on areas with more growth. That is a category exit by a corporate parent, eleven months ago. And flowerbeauty.com is trading today, with a full catalogue and a store locator pointing at CVS and Ulta.

Which leaves Gxve Beauty, where the honest answer is that nobody knows. Cosmetics Business ran its report under a question mark, used the words allegedly and rumoured, noted that Gwen Stefani had not commented, and recorded that its email to the brand bounced. The domains no longer resolve, which is suggestive. No official announcement exists from anyone, and a rumour repeated confidently is still a rumour, however deep into the thicket it has travelled.

Why the miscount matters#

This is not pedantry about a trade article. It is about what happens when a marketing team builds a celebrity brands strategy on a list it has not walked itself. Five simultaneous celebrity brands collapsing in eighteen months reads as a category in crisis, and a category in crisis produces a defensive brief: avoid the space, cut the ambition, do not launch.

One clean closure, two 2025 restructurings, one unconfirmed and one that has not happened reads as something entirely different. It reads as an ordinary distribution of outcomes in a hard category, which invites a much better question than "are celebrity brands dead". The better question is which ones survive, and why. That is a question with an answer, and the answer has been sitting in the hedgerow of the academic record for twenty-five years.

What the celebrity endorsement research actually found#

The useful evidence is not a list of brands at all. It is a meta-analysis. Knoll and Matthes, in the Journal of the Academy of Marketing Science in 2016, pooled 46 studies covering 10,357 participants and found that the effect of celebrity endorsement does not merely vary in size. It flips sign.

In the authors' own reporting, the most positive attitudinal effect appeared for male actors matched well with an implicitly endorsed object, at d = .90, and the most negative for female models matched poorly with an explicitly endorsed object, at d = -.96. Both are large effects. The variable separating them is fit, not fame, and the paper reports no publication bias.

There is a second finding in the same paper that ought to be quoted far more often than it is: celebrity endorsements performed worse than endorsements from quality seals, awards or endorser brands. A trust mark beat a famous face. For any category where the buyer is nervous, and folkfox works almost entirely in categories where the buyer is nervous, that is the whole strategy in one sentence. The certification outruns the celebrity, every time, in the places where trust is the thing actually being bought.

None of this is new. Till and Busler established the match-up hypothesis a quarter of a century ago, finding that fit, or belongingness, plays an important role in whether an endorsement lands. Völckner and Sattler reached the same conclusion from the other direction, finding that fit between a parent brand and an extension product is the most important driver of extension success. Three separate literatures, one answer.

Where permission is easy, and where it is bought#

Fit is not evenly distributed across categories, and there is consumer data that puts numbers on where celebrity brands still buy their way in. A panel survey published by Highlight of 1,296 panel members, fielded in late October 2025, asked where a celebrity association actually moves a purchase decision.

Line that up against the five brands and the pattern is almost too neat to trust, so treat it as a prompt rather than a proof. Skkn, Gxve and Flower are beauty, the category where a celebrity association is strongest and where all three ended in restructuring or ambiguity rather than outright market failure. Más+ and Unwell are drinks, at 36 per cent, where permission is much harder to borrow. The one clean closure sits in the harder category.

The same survey found 60 per cent of respondents sceptical of celebrity-backed products generally, 44 per cent assuming the celebrity had simply been paid to market it, and only 18 per cent believing the celebrity was heavily involved in developing it. Treat those as panel figures rather than population estimates, but the direction is consistent with the meta-analysis: the default assumption is absence, and fit is what overturns it.

Which is why the brief that starts with reach is already lost. Reach tells you how many people will see the launch. It says nothing about how many will grant the founder standing in that aisle, and standing is the thing being tested. A brand can be seen by everyone and believed by nobody, and the numbers above are what that looks like before the first case ships. Most celebrity brands that fail did not fail at awareness.

What the most successful celebrity brands get right#

So, what brands are owned by celebrities and actually work? The pattern in the research and in the survivors is consistent enough to state plainly. The most successful celebrity brands are the ones where the founder has a credible reason to be in that aisle, where the product would survive a blind test against its shelf neighbours, and where distribution was built rather than announced.

Which turns the brief on its head. The question is not whether a famous founder helps. It is whether the category will let them in, which is the only question about celebrity brands worth paying for, and that is answerable before a penny is spent, using the same three checks that corrected the list at the top of this piece: what does the company itself say, when did it happen, and is anyone still buying.

The quiet quarry here is not fame at all. It is fit, patiently established, and the discipline to check a claim before building a plan on it. Attention is loud and cheap and arrives on its own, a whole clamour of it, lanterns in the dark. Permission is earned on a patient prowl through the undergrowth, one credible product at a time, and no amount of reach substitutes for it. A fox that mistakes noise for a meal goes hungry.

Frequently asked questions#