Ninety four percent put it back

Listen to this episode of The Daily Trend Report, then use the piece below.

The Food Dive piece going around about nutrition scanner apps getting your product wrong is not a news story. It ran under Opinion on September 24, and the byline belongs to Cara Liebrock, CEO of Perfect Snacks. Perfect Bar is sweetened with organic honey. The piece argues, at some length, that counting honey as "added sugar" is one of the ways these scoring algorithms miss.

She's right, by the way. That's the part people will skip on the way to the gotcha, and skipping it is the mistake.

Honey and maple syrup do land in the added sugars line the same as refined sugar. Plain yogurt and whole grain bread are both processed foods by any working definition. A score that treats those as strikes is a score with a resolution problem, and if you make real food you have felt it. Liebrock calls the fix "food fluency" and puts the job on brands to teach it. She cites a Harris Poll run for Perfect Bar saying 63% of Americans avoid products with ingredients they don't recognize.

So: a correct argument, a coined term, a commissioned stat, placed in the trade publication the buyers read, under the name of the person who runs the brand it helps most.

That is not a scandal. That is the play.

Perfect Snacks has been a Mondelez brand since 2019. Food Dive's own reporting last November put the brand at $156.4 million in trailing sales, up about 20%, on Nielsen numbers the company supplied, with Liebrock telling Christopher Doering she'd been "very clear with Mondelez that we're sitting on a $1 billion brand." A brand at that size does not argue with a scanner app. It goes over the scanner app's head and argues with the standard the scanner app is judged by, in front of the buyers who set shelf.

Now here is the number that decides whether any of this matters to you, and it was published seven months ago by somebody else.

Ninety four percent of Yuka's US users say they will put a product back on the shelf if it gets a bad rating. That is Yuka's own research, reported by Harry Holmes at FoodNavigator on February 9. Same piece: Yuka is past 80 million users across 14 countries, and 92% of US users say they're buying fewer ultra-processed products since they started scanning.

Ninety four percent. Not "shoppers are increasingly influenced by." Ninety four out of a hundred, self-reported, standing in the aisle, holding the thing.

The ranking is as strange as you'd expect. Holmes reports Alpro's unsweetened oat milk at 78 out of 100, rated excellent, while the unsweetened soy milk sits at 49 and rates poor. Proper Chips, a UK crisp built specifically as the better option, scores under Doritos and Walkers. Most of a Yuka score runs on Nutri-Score, which rewards protein and fiber and penalizes calories, sugar, salt and saturated fat, so a small, fatty, honest product gets read as a problem and a reformulated one that hits the macros walks.

We talked about this on the show this morning and I said the thing everybody says, that there's nobody to call. No appeals desk, no ticket, version numbers you'll never see. That's true about appeals and it's the wrong place to be looking, and I want to correct it here because it changes what you do on Tuesday.

The traffic already runs the other direction. Since 2024, a Yuka user who scans a product with what the app flags as a high-risk additive can send a message straight to the brand asking them to pull it. Holmes reports Chobani took dipotassium phosphate out of its oat milk after getting those, and that the average count of high-risk additives per product in Yuka's France database is down 13% since 2019. Yuka's US country manager, Gabriella Sebag-Weingrad, says the goal is exactly that, moving manufacturers faster than regulation would.

So there's no desk to call, but there is an inbox, and it's yours. Somebody has already written to you about an ingredient. You just filed it under complaints.

Put those two things next to each other and the shape of the week gets clear. Chasing the number is the expensive move and you will lose it, because a conglomerate can run reformulation against a scoring model all year and you have one co-packer window and one shot at a spec change. Changing the frame is the cheap move, and Liebrock just showed you what it looks like at the top of the market. You can't commission a Harris Poll. You can publish the one thing about your product that no scoring model has a field for, on camera, this week, for nothing. Same play, different budget.

Even Segun Akinwoleola, who runs Gym Kitchen in the UK and whose ready meals score well, told Holmes the single score has limits and pushes a simple good-versus-bad framing. That's a brand winning under the system saying the system is thin. You are allowed to say it too.

Three moves, all free.

One, the pasteable post. Scan your own product, screen record it, and publish the number whatever it is. Three rules. Show the actual score, including a bad one, because a brand that posts its own mediocre rating and then explains it reads honest in a way no caption buys. Do not argue with the model like a lawyer and do not name the app as an enemy, since that turns your post into a fight nobody asked you to pick. Say one specific thing the score cannot see, and make it a fact with a name attached, not an adjective. "The fruit came in over three weeks in August and we run it the day it lands" beats "made with care" every time, because the first one you can be caught lying about.

Two, the still. Your hand, your phone, your product's score on the screen, shot in front of the shelf where your product actually sits, in a store that really carries you. Not a studio and not a nicer store. Leave the other brands in the frame. Leave the shelf tag in. The reason this works is that it is provably the aisle, and a flat lay on white seamless is provably not.

Three, the video, fifteen to thirty seconds, five shots, phone, no captions burned in. Shot one, wide, you walking up to your own shelf. Shot two, close, your thumb opening the app. Shot three, the scan and the score, held long enough to read, unedited. Shot four, tight on the product in your hand while you say the one thing the model has no field for. Shot five, you put it in the cart and walk. Two rules. Say the full product name out loud in the first five seconds and again at the end, because search is the one distribution route nobody can outbid you for. And do not reshoot the scan to get a better number.

The free internal one, and it takes twenty minutes. Open your support inbox and your DMs, search the last twelve months for the names of the four or five additives in your own ingredient panel, and count the hits. Those messages are a free panel study somebody else paid to run on your product, sorted by the exact thing the scanner flagged. If the count is zero, good, now you know the scanner is not your problem and you can stop reading about it. If the count is not zero, you have your content calendar and it wrote itself.

One caution, because this only works one way. Do not run any of it if the score is simply correct about you. Filming is not a laundry service. It works because you are telling the truth on camera, and if the truth is that the product is what the app says it is, the camera makes that worse, faster, in front of more people.

The show this morning went through this with Chloe Dawn and Rachel Donovan, along with Pinterest making the front row phone-free at London Fashion Week and why Knix put a prototype count in a press release. The episode is the briefing. This is the extension. Subscribe on Apple Podcasts or on Spotify and it's in your feed before the store opens.

Pick your three best sellers, scan them tonight in whatever your buyers are actually using, and find out what your shelf already says about you. Somebody is scanning you this month either way. The only open question is whether your version of that video got there first.

I do this for a handful of food and beverage brands a quarter. If yours is next, find me on LinkedIn.