Supplement de-influencing: why we built a tool to check what TikTok won't tell you

The #deinfluencing hashtag crossed 1 billion views on TikTok in 2023 and has since passed 3.5 billion total content views. It started as people saying "don't buy this", and it was refreshing. After years of #TikTokMadeMeBuyIt hauls, someone finally said the quiet thing out loud: most of this stuff isn't worth your money.

There is a problem with it, though.

De-influencing is still influencing. The format changed. The incentives did not. A creator telling you not to buy AG1 while recommending their own code for a different greens powder is not transparency. It is marketing with a sceptical coat of paint. Research from MediaCat UK found exactly this: de-influencing videos build trust in the creator, but do not change viewers' purchasing behaviour or brand perceptions.

Meanwhile, 74% of Gen Z users now think twice about who they engage with on TikTok, and 79% say they miss the pre-commercialised version of the platform. Trust is falling. The supplement recommendations keep coming.

I did not build De-Influenced because I wanted to be a content creator. I built it because I needed it.

Why I built this

I'm Liz. I have worked in healthcare IT for over 20 years, and my partner works in frontline healthcare. Between us, we have seen how the gap between clinical evidence and consumer marketing plays out, and I have spent more money than I would like to admit on supplements over the past decade.

Some of them were useful. Many were not. The problem was always the same: I could not tell which was which by looking at the label.

I would see a product with 15 ingredients, four of them inside a "proprietary blend" with no individual doses, marketed by someone with a ring light and a discount code. I would Google the product, and every review was either an affiliate post or a personal anecdote. Nobody was checking whether the ingredient doses matched the clinical research. Nobody was saying "this contains 50mg of ashwagandha, but the studies used 600mg."

So I started doing it myself. Manually. Pulling up labels, cross-referencing PubMed, checking clinical trials. Then I thought: why isn't there a tool that does this?

That is what De-Influenced is. You paste a product URL and upload a photo of the ingredient label. The tool extracts the ingredients, checks every dose against clinical evidence, scores the label for transparency, and shows alternatives that are cheaper and properly dosed. No affiliate links. No sponsored placements. No discount codes.

It exists because I needed it. I suspect you do too.

Three TikTok supplement trends that show the problem

"Nature's Ozempic": the berberine myth

Berberine is currently the #1 selling supplement on UK TikTok Shop, with nearly 5,000 orders a week and over £815,000 in weekly revenue. It is marketed as "nature's Ozempic", a natural alternative to semaglutide, the prescription GLP-1 drug that produces large weight loss.

The clinical evidence says something different.

A 2026 randomised controlled trial in JAMA Network Open, with 337 participants, double-blind and placebo-controlled across 11 hospitals, found that berberine at 1g daily for six months did not reduce visceral fat or liver fat compared with placebo.

A 2025 systematic review of 23 RCTs found berberine reduced body weight by an average of 0.88kg and BMI by 0.48 points. For context, semaglutide (Wegovy) produces about 14.9% body weight loss at 68 weeks, roughly 15kg for someone starting at 100kg.

These are not comparable outcomes. As the American Academy of Family Physicians noted, berberine "lacks rigorous evidence and has potential harms." Those harms include documented interactions with blood thinners, immunosuppressants and diabetes medication, none of which appear on the TikTok Shop product pages I have checked.

Berberine has real research for blood sugar management in people with metabolic conditions. That is a genuine use. "Nature's Ozempic" is not a clinical description. It is a line written to move units.

Collagen gummies: the dosing problem nobody mentions

Collagen supplements are a £200 million-plus UK market, and gummies are the fastest-growing format. The appeal is obvious. They taste good and they are easy.

The problem is the dose.

Clinical trials showing benefits for skin hydration and elasticity consistently use hydrolysed collagen peptides at 2.5 to 10g a day. A 2024 RCT in the Journal of Clinical Medicine used 10g daily and found improvements in skin elasticity and hydration against placebo. A systematic review of 26 RCTs covering 1,721 patients confirmed that hydrolysed collagen improved skin hydration and elasticity, at doses of 2.5g to 10g daily.

Most collagen gummies contain 1 to 3g a serving. Some contain less. At the lower end you are getting a fraction of the dose the research tested. The format is part of the problem: you cannot fit 10g of collagen peptides into two gummy sweets without them being the size of golf balls.

This does not mean collagen is useless. It means the format matters, the dose matters, and "clinically studied ingredient" on the label does not mean "clinically studied dose in this product."

"Cortisol face" and the ashwagandha rush

"Cortisol face" is not a medical diagnosis. It is a TikTok trend that claims everyday stress visibly puffs up your face, and that the fix is ashwagandha to lower cortisol.

Healthline reported what dermatologists and endocrinologists have confirmed: the facial swelling associated with high cortisol, called "moon face", is a symptom of Cushing's syndrome, which affects about 10 to 15 people per million. Normal everyday stress does not cause it.

As Dr Gary Goldenberg, a board-certified New York dermatologist, told Prevention: "Stress alone cannot cause cortisol face puffiness. This is typically caused by a medical condition."

Ashwagandha itself does have research behind it. A 2012 study in the Indian Journal of Psychological Medicine found that 300mg twice daily of KSM-66 ashwagandha root extract reduced serum cortisol by 27.9% against 7.9% in the placebo group. A 2019 study in Cureus found that 600mg daily reduced cortisol and improved sleep quality. Those are real findings from properly run trials.

The detail that matters is this: the studies used 300 to 600mg of a specific standardised extract (KSM-66), measured over 8 to 12 weeks. Most TikTok-promoted ashwagandha products either do not say which extract, do not say the dose, or bury it in a proprietary blend where the amount could be anything from 50mg to 500mg. You cannot tell whether you are getting a meaningful dose.

Even when the ashwagandha is properly dosed, it is being sold as a fix for "cortisol face", which is not a real condition. That is a solution to a problem that does not exist, sold to people a 30-second video convinced they have it.

What de-influencing gets right, and what it misses

The de-influencing movement diagnosed a real problem: people are being sold overpriced, underdosed supplements by creators who are paid to recommend them. That is accurate and it matters.

What it misses is the solution.

Telling people "don't buy this" without explaining why, without checking the dose, citing the research or comparing the alternatives, is opinion dressed up as advice. It builds the creator's credibility. It does not build the viewer's knowledge.

The supplement industry does not need more opinions. It needs more transparency, and tools that let consumers verify claims themselves, without trusting any single creator, brand or reviewer.

That is what De-Influenced does. Not "trust me instead of them." Here is the label, here is the research, here is what the dose is, and here are the alternatives.

How De-Influenced works

1. Paste a product URL and upload the ingredient label. Any supplement sold online in the UK.
2. We extract the label. Every ingredient, every dose, every blend.
3. We check it. Doses are compared with published clinical research. Ingredient forms are scored (magnesium glycinate against magnesium oxide, for example). Proprietary blends are flagged.
4. You get a transparency score. A clear number for how much the label tells you. AG1's UK label, for example, prints a dose for only 31 of its 73 active ingredients.
5. We show alternatives. Products with comparable ingredients at clinical doses, sorted by price and quality tier.

No affiliate links. No sponsored content. No brand partnerships. Your first analysis is free at De-Influenced.

---

Frequently asked questions

What is supplement de-influencing?

Supplement de-influencing is a social media movement where creators advise against buying specific supplements, typically products promoted by influencers or through paid partnerships. The #deinfluencing hashtag has passed 3.5 billion views on TikTok. The intent is to counter overhyped marketing, but the format still rests on one creator's opinion rather than systematic, evidence-based analysis.

Is berberine actually "nature's Ozempic"?

No. Semaglutide (Ozempic and Wegovy) produces about 14.9% body weight loss in trials. A 2025 meta-analysis of 23 RCTs found berberine reduces body weight by an average of 0.88kg. A 2026 JAMA Network Open RCT found berberine did not reduce visceral fat against placebo. Berberine has research for blood sugar regulation in metabolic conditions, but it is not comparable to GLP-1 drugs for weight loss.

How much collagen do you need for skin benefits?

Trials showing improvements in skin hydration and elasticity used hydrolysed collagen peptides at 2.5 to 10g a day. Most collagen gummies provide 1 to 3g a serving. Powder or liquid formats are more likely to deliver a meaningful dose.

Does ashwagandha lower cortisol?

Some evidence supports this. A 2019 study in Cureus found 600mg daily of ashwagandha root extract reduced serum cortisol against placebo. A 2012 study found similar results at 300mg twice daily of KSM-66 extract, a 27.9% cortisol reduction. Not all ashwagandha products contain the same extract, dose or standardisation, and "cortisol face" is not a recognised medical condition.

Is De-Influenced free?

Your first analysis is free. A subscription unlocks 50 analyses a month and the alternatives. No affiliate links, no sponsored content.

Who is behind De-Influenced?

De-Influenced was founded by Liz Whittaker, with over 20 years in healthcare IT and a partner working in frontline healthcare. The tool was built after years of researching supplement labels and cross-referencing clinical evidence by hand. It exists to give consumers the same analysis without reading PubMed at midnight.

---

Medical disclaimer: This article is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional before starting, stopping or changing any supplement regimen, particularly if you take prescription medication, are pregnant, or have an existing health condition. De-Influenced is a supplement transparency tool, not a medical service.