Perfect Corp. (PERF)

Perfect Corp. builds software that lets people try on makeup and beauty products on their phones before buying them. Think of it as a mirror — but digital, interactive, and connected to a shopping experience. The company’s technology uses artificial intelligence and computer vision to map a person’s face, simulate how different products look on that specific face, and then link the experience directly to retail, so a customer can buy what they just tried. It is used by hundreds of beauty brands and retailers around the world, including major cosmetics companies and online sellers.

The problem and the product

Shopping for makeup is hard. You walk into a store or scroll through an online catalog and see a shade or finish, but you cannot really know if it will look good on you until you try it. Buying the wrong thing is annoying and wasteful. Returns are a hassle. Some people are shy about testing in-store, and many times you cannot test at all when ordering online.

Perfect Corp. solved this with software. Their main app, YouCam Makeup, uses your phone’s camera and AI to detect your face, then overlay makeup products onto your image in real time. You can see how a lipstick shade looks, how an eyeshadow blends, how a foundation matches your skin tone. All before you buy. The AI learns your face’s shape and coloring, and the simulation is detailed enough to be useful — not perfect, but close enough to guide a purchase.

The company licensed this technology to beauty brands and retailers. If you visit Estée Lauder’s website or a Sephora app, there is a good chance you are using Perfect Corp.’s technology behind the scenes. The company makes money by licensing the software, selling ads inside the app, and taking a small cut of sales that happen through the virtual try-on experience.

Why this matters to beauty companies

For a cosmetics brand, virtual try-on is valuable. It reduces return rates — fewer people buy the wrong shade if they tried it first. It increases customer confidence and willingness to buy online, where margins are often better than in physical retail. It collects data: the brands learn which shades and finishes are being tried most, which ones convert to sales, and what people are searching for. That data shapes product development and marketing.

For retailers, the same logic applies. Sephora and Ulta and other beauty retailers want to reduce returns and increase online conversion. They also want the technology because it is a draw — something that makes their app or website better than a competitor’s. Perfect Corp. is a supplier to both brands and retailers, sitting in the middle of the beauty e-commerce ecosystem.

The scale of the business

Perfect Corp. generated revenue in the low hundreds of millions in recent years, though the company is not uniformly profitable. Most of the money comes from licensing deals with major brands and retailers, though the company also runs its own consumer apps — like YouCam Makeup as a standalone mobile app — which generate revenue through subscriptions and in-app purchases.

The business model is software-as-a-service, which means revenue is often recurring: a beauty brand pays a subscription to use the technology, and that contract renews yearly or with a multi-year commitment. That recurring revenue is valuable because it is predictable and it compounds over time. But it also means the company has to keep customers happy and keep its technology competitive, or they will switch to a rival.

The competitive landscape

Perfect Corp. does not compete against a single rival. Instead, it competes in a few overlapping spaces. First, there are other virtual-try-on technology companies building similar AI and computer-vision tools for beauty. Modiface, owned by Loreal, is one example — a company with a major beauty conglomerate backing it, which gives it scale and resources. Second, Perfect Corp. competes against the do-it-yourself approach: larger brands like Loreal or Estée Lauder could build try-on technology in-house rather than licensing from Perfect Corp. Third, it competes against the in-person experience — the reason you might try on makeup at a store counter instead of using an app.

The company’s strength is that it has been building this technology since 2011, before virtual try-on was mainstream. It has accumulated a large number of brand and retailer customers, trained its AI on millions of faces and cosmetic products, and built integrations deep into major retail platforms. That head start matters in a software business because switching costs are real — once a brand has trained staff on how to use the platform and customers expect a certain experience, changing to a new vendor is disruptive.

The weakness is that capital-rich competitors could build or buy competing technology, and that the beauty industry is concentrated among a few very large groups — Loreal, Estée Lauder, Coty, and others — who have the resources to build in-house if they chose to.

The business dynamics

Perfect Corp. has grown revenue over time, but profitability has been elusive. The company invests heavily in research and development to keep its AI models current, in sales and partnerships to win new brand clients, and in product teams to keep the consumer apps competitive. Those investments are necessary but they eat into margins.

The company also faces a conversion problem: not everyone who tries on makeup buys. The virtual try-on experience is useful, but it is not the same as holding a product in your hand. Some of the value to customers is free trial without purchase. For the company to be profitable, enough users have to convert to sales, and the take rate on those sales has to be enough to cover the cost of building and maintaining the technology.

What to watch

Someone researching Perfect Corp. should read the 10-K (SEC CIK 0001899830) to understand the revenue mix — how much comes from each brand or retailer, what the largest customer is, and whether any single customer is too large a share of revenue. Customer concentration matters because if one large client leaves, revenue can drop fast.

The other key number is the gross margin on the licensing business: the take-rate after paying infrastructure costs and customer support. A healthy gross margin means the company is extracting real value from its technology. A shrinking gross margin means competitors are pushing prices down or customers are negotiating harder. Watch for trends in the number of active brand clients and the contract value of new deals compared to existing ones. Growing customer count and bigger deals are signs of momentum; flat or declining metrics suggest trouble.

The profitability question is fundamental. As the market for virtual try-on matures, will Perfect Corp. be able to improve margins, or will competition and customer pressure keep squeezing them? That is the bet.