Nine beauty brands run virtual try-on in our database of 68 AR commerce deployments, last verified on September 15, 2026. Two vendors power seven of the nine. Only five of the nine publish any result at all, and the column this database keeps for return rates is empty for every brand in it, beauty included. This post counts the AR beauty try-on market as recorded: who runs it, who supplies the technology, and how little of the promised effect has ever been put in writing.
The nine brands, and the two vendors behind seven of them
Start with the supply side. ModiFace, the try-on specialist owned by L'Oréal, powers five of the nine recorded beauty deployments: L'Oréal itself, Sephora, Estée Lauder, Maybelline and NYX. Perfect Corp powers two more, MAC and Ulta Beauty. The remaining two sit outside the duopoly: Charlotte Tilbury runs its Magic Mirror with Holition and its 3D virtual store with Obsess, and Nykaa built its try-on in-house. Every recorded beauty launch in the database falls between 2016 and 2022, which makes MAC, the most recent at 2022, the newest beauty try-on launch we can verify.
| Brand | Vendor | Launched | Publishes a result? |
|---|---|---|---|
| L'Oréal | ModiFace | 2018 | Yes: conversion and engagement |
| Sephora | ModiFace | 2016 | Yes: engagement |
| Estée Lauder | ModiFace | 2017 | Yes: conversion (vendor claim) |
| Maybelline | ModiFace | 2019 | Nothing published |
| NYX Professional Makeup | ModiFace | 2019 | Nothing published |
| MAC Cosmetics | Perfect Corp | 2022 | Yes: conversion and engagement |
| Ulta Beauty | Perfect Corp | 2016 | Yes: engagement |
| Charlotte Tilbury | Holition and Obsess | 2020 | Nothing published |
| Nykaa | In-house | Not recorded | Nothing published |
Concentration like this cuts both ways. A retailer choosing a beauty try-on vendor today is really choosing between two proven suppliers and a long tail, and when one of the two changes its product, five to seven brand experiences change with it. What concentration does not tell you is whether any of it works. For that, you need the published results, and they are much thinner than the deployment count suggests.
The five published results, labeled for what they are
Every figure in this database is a brand-reported or vendor-reported marketing figure, not an independently tested measurement, and each one below carries its source so the label stays attached. Five of the nine beauty brands publish something. L'Oréal reports a 200% conversion lift and a 100% engagement lift (reported by The Drum). MAC reports 62% on conversion and 200% on engagement (Adweek). Estée Lauder's 84% conversion figure is a vendor claim, quoted from a ModiFace executive in Cosmetics Business, and our own dataset notes flag it as such. Ulta Beauty reports a 600% engagement lift (Harvard Business Review, via GCI Magazine). Sephora reports a 16% engagement lift (a Braze case study). That is the entire published evidence base for beauty try-on in the record: five brands, eight numbers, none of them audited.
The other four publish nothing at all. Maybelline and NYX have sold ModiFace try-on on Amazon and their own sites since 2019 without a single results figure entering the record. Charlotte Tilbury and Nykaa are equally silent. Silence is not proof a deployment failed; brands publish numbers that flatter them and sit on the rest. But it does mean the category's reputation rests on the five who chose to talk.

Beauty is not unusually quiet, either. Across all 68 brands in the database, only 12 publish any lift figure, and only 8 publish a conversion lift. Those eight have a median of 73%, a median built from eight self-reported numbers, so treat it as a measure of what brands claim, not of what try-on does. The neighboring category shows the same pattern in miniature in our count of eyewear brands and virtual try-on conversion, where six brands produced two sourced conversion lifts, Eyebuydirect at 30% and Lenskart at 17%.
The empty column: returns
Fewer returns is the promise attached to almost every try-on pitch: match the shade on your own face first, and the wrong shade never ships. Our database keeps a return-rate column for exactly this claim, one row per brand. For beauty, it is empty nine times out of nine. For the full database, it is empty 68 times out of 68. The reason is method, not oversight: a figure is recorded only when a published, attributable source exists, and no brand-level AR return-rate figure from a vendor case study or brand announcement was found for any brand. Even the most repeated number in this niche, the claim that Wayfair cut returns by 43% with AR, has no vendor case study or brand press release behind it, so it was deliberately left blank. The honest answer to whether virtual try-on reduces returns for makeup is that the brands selling it have not published the number, a gap we work through in Does Augmented Reality Reduce Ecommerce Returns?
The newest move arrived on October 29, and it published no number either
On October 29, 2026, Maybelline announced in its own launch release that it had become the first beauty brand to bring real-time makeup virtual try-on into ChatGPT, powered by ModiFace, with shade advice and try-on inside the chat. The channel is genuinely new: try-on is leaving brand product pages for the chat window. The evidence habit is not new at all. Maybelline remains one of the four recorded beauty brands that has never published a try-on result, so the biggest distribution move in the category's history arrives with reach claims and, so far, no results claims.
What this count does and does not say, and one photo tool that is not try-on
The method, plainly: 68 brands with verifiable 3D or AR on their product pages, checked on September 15, 2026, with statistics taken only from published sources and blanks left where no attributable figure exists. The count says the beauty try-on supply is concentrated and its published evidence is thin. It does not say try-on fails; it says the public record cannot currently prove it works, which is a different and more useful thing to know before signing a vendor contract. For the results themselves, brand by brand and story by story, read Beauty Brands With Virtual Try-On: Results Worth Noting, the narrative this census counts underneath.
One photo-based tool on this site works the opposite end of the same product photo, and it is worth keeping the two apart. The 2D to 3D Photo Converter turns one photo into a printable relief or lithophane, a physical object a seller can hold. It does not do try-on and it cannot show a shade on a face. Try-on shows the product on you before you buy; the converter makes a thing out of the photo after you have it. Both start from a single photo, and that is where the resemblance ends.