Virtual Makeup Try-On for Event Activations: A Booth That Works Without Testers

A beauty trade show, a pop-up in a shopping mall, a launch activation at an influencer event: the script is usually the same. A table with displays, open testers, cotton pads, alcohol, a brand ambassador wiping a lipstick bullet between one person and the next, and a queue that moves slowly. Virtual makeup try-on for event activations changes that mechanic: instead of applying the product to the visitor's skin, the brand shows the colour rendered on her face, in the browser, in seconds — and frees the team to talk instead of sanitising.

This article is a practical set-up guide: what the technology solves, what it does not solve, which decisions need to be made with the agency before the event and how the data generated at the booth flows back into the brand.

What a tester costs in an activation

A tester is not just a line in the materials budget. It sets the pace of the booth.

Every real application consumes ambassador time, requires sanitising, limits how many colours the visitor is willing to try and creates a physical bottleneck: while one person tests, the others wait. In activations with a wide range — a lipstick collection, a foundation line with many shades — the visitor tries two or three colours and leaves without seeing the rest.

There is also the post-pandemic effect: part of the public simply does not want a shared product on their face. And there is the replenishment problem: testers run out, break, go missing. By the second day of the show the booth no longer displays the full range.

Virtual makeup try-on for event activations: what runs at the booth

MaIA Try-On exists in two forms, and the choice between them changes the design of the activation.

Live mode through the camera

The browser SDK, in early access since September 2026, covers thirteen products — lipstick, foundation, blush, face powder, under-eye concealer, eyeshadow, mascara, eyebrow makeup, eyeliner, bronzer, contour, highlighter and lip liner — plus a combined Full Look. In Live mode, rendering happens locally in the browser, using the device's camera: nothing is sent to a server. The visitor sees herself moving, turns her head, switches colours.

For a booth, that solves three things at once: the queue moves faster, the full range is available on one screen and the privacy conversation becomes simple — there is no image upload.

The SDK is self-hosted by the brand and released per account, in early access. It is not a public download and it does not come with device certification, so the booth hardware has to be tested beforehand, with the browser and the camera that will actually be used on the day.

Photo mode and the API

The SDK's Photo mode and the try-on API work from an image: the visitor takes a photo, the product is rendered and the image comes back. The API covers lipstick and foundation, with palettes calibrated to the brand's real products inside a project.

This format is the one that serves "take your photo home": the visitor leaves the booth with an image to post, or receives the result later by WhatsApp or e-mail. If you are unsure which path suits your activation, it is worth reading the comparison between photo-based try-on and live camera try-on.

How to design the booth flow

A format that works well is to split the activation into three short stations.

Station 1 — discovery

A selfie feeds the profile analysis: skin tone and undertone, apparent skin type and condition, hair attributes, with a confidence score per field. That is what skin and hair analysis from a selfie delivers as JSON. Here the ambassador stops asking "are you fair or medium?" and starts showing a reading. Important: what appears are visible attributes of the photo, not a clinical assessment — nobody at the booth should talk about treatment or a clinical condition.

Station 2 — try-on

With the profile in hand, the screen already opens with the most likely colours highlighted, but with the full range within reach. This is where Live mode comes in: the visitor switches lipstick, blush and eyeshadow on her own, and the ambassador steps in only to guide. The Full Look helps when the activation is launching a collection that was designed as a set.

Station 3 — conversion

The visitor leaves with a list of SKUs, a coupon or a link. In activations with no stock on site, this station is what turns experience into an order: the very colour she tried becomes an item in the e-commerce cart.

Technical agreements before the event

A few points need to be settled with the agency and with the booth supplier:

• Lighting. A hard spotlight from above ruins the photo and gets in the way of reading the tone. Diffuse, frontal light is the safe bet.
• Connectivity. Convention-centre Wi-Fi drops. Agree on a 4G/5G plan B and test it on set-up day, not on the eve of opening.
• Devices. Define model, browser and version, and run the camera test on the actual hardware.
• Colours. The SDK accepts any colour as #RRGGBB; someone at the brand has to approve the match between each hex and each SKU before the event.
• Consent. If the activation captures a photo or a contact, the consent text has to be visible on screen and on the sign-up form, with a clear purpose.

The data that remains after the event

The biggest loss of a traditional activation is that it ends when the booth is dismantled. With a beauty profile collected under consent, the brand leaves the event with something beyond a visitor count: tones, undertones, colours tried and colours abandoned.

That feeds the CRM, the e-mail cadence and the next conversation — including with a virtual beauty advisor integrated with the catalog, which picks the contact back up knowing what the person tried. Results vary with each brand's catalog, traffic and channel, so treat the first event as calibration: record what converted and adjust the flow at the next activation.

Want to test the flow before the next show? Create your account and API key in the MaIA API documentation and run the first requests, or talk to the team to evaluate early access to the SDK.