Beauty Profile Product Recommendations for Out-of-Stock Items: The Right Substitute Instead of a "Notify Me"

Stockouts in beauty come with an extra complication: having another item in the same category isn't enough, it has to work on that particular skin. That's why beauty profile product recommendations for out-of-stock items are one of the most direct applications of a structured profile: instead of pushing the shopper toward a "notify me when it's back" button, the store offers the closest available SKU to what she was looking for — and explains why.

What happens today when a shade sells out

In most beauty stores, the page for a sold-out shade does three things: it disables the buy button, opens an email field and, sometimes, shows a "related products" carousel built on popularity or category. None of the three understands that the shopper's problem isn't "lipstick", it's "that red with a cool undertone that suits me".

The outcome is predictable. Part of the traffic leaves the session. Part signs up and comes back weeks later, if at all. And part buys a neighboring shade blind, which usually turns into an exchange or a return — exactly the cost the operation is trying to avoid.

The right substitute is an attribute problem

A good substitute is rarely "the next number on the scale". It's the product whose descriptive attributes come closest: shade range, undertone, finish, coverage, texture. When the catalog only holds a commercial name and a color code, there's no way to calculate that proximity. When it holds attributes, substitution stops being guesswork and becomes a rule.

Beauty profile product recommendations for out-of-stock items in practice

The logic is simple to describe and demands data discipline. Instead of linking a sold-out SKU to an alternative SKU in a manual table, you link the shopper to a profile and the profile to the available catalog.

Step 1: capture the profile

MaIA Scan receives a selfie via API and returns a JSON with visible attributes: skin tone and undertone, skin type and condition, hair attributes and a confidence score per field. These are attributes observable in the photo — a descriptive reading, not a medical assessment, and it does not replace a professional.

On the page of an out-of-stock item, that invitation has a clear reason for the shopper: "this shade is sold out; send a selfie and we'll find the closest one available". It's a far more concrete value exchange than a notification form.

Step 2: describe the catalog with the same attributes

The profile is only useful if the catalog speaks the same language. That means enriching SKUs with shade range, undertone, finish and coverage — and keeping that enrichment current with every collection. It's the least glamorous part of the project and the one that defines the quality of every recommendation, not just the stockout ones.

Step 3: filter by availability before ranking

The golden rule of substitution is to invert the usual order: first filter what's in stock at the distribution center serving that postal code, then rank by attribute proximity to the sold-out item and to the profile. Precisely recommending a product that will also run out is worse than not recommending at all.

Step 4: explain the choice

"This shade has the same warm undertone and slightly lighter coverage" converts better than "you might also like". The justification sentence can be generated from the profile fields and the SKU attributes themselves, with no hand-written copy per product.

Let the shopper see before she switches

Substitution carries perceived risk: the person wanted a specific item and is being invited to accept another one. A preview reduces that friction. With virtual makeup try-on, the API renders lipstick or foundation over the submitted photo, with palettes calibrated to the brand's real products in the project. The browser SDK, in early access, covers thirteen products plus the Full Look in live camera mode or photo mode, rendered locally on the shopper's machine.

In practice, the sequence becomes: sold-out shade → profile → two or three available substitutes → preview of the preferred substitute → cart. Each step answers a different objection.

Where the same logic repeats

Beyond the product page

Stockouts don't only happen on the PDP. They show up in carts abandoned for lack of stock, in site search returning unavailable results and in the replenishment flow, when a shopper comes back for the item she always uses and it has vanished from the catalog. At all of those points, the profile already captured lets you respond with a substitute instead of an empty space.

In customer service and on WhatsApp

When conversation is the channel, the virtual beauty advisor does this work inside the dialogue: it reads the profile, checks the brand's catalog and proposes the available alternative, with the rationale in natural language. It's a project run with the B4A team, integrated with the brand's catalog and stock rules.

In permanent discontinuation

The most sensitive case is the item that isn't coming back. There, attribute-based substitution stops being a temporary workaround and becomes the only possible answer — and an opportunity to migrate the customer base from a retired shade to the right successor, instead of losing it to a competitor.

What to measure

Track the click-through rate on suggested substitutes on out-of-stock pages, the conversion of those sessions compared with the "notify me" flow, the return rate of orders originated by substitution and the percentage of profiles captured on sold-out PDPs. Results vary with each brand's catalog, traffic and channel — treat the first few months as calibration, not as a verdict.

Want to test attribute-based substitution on your catalog? Create your account and API key in the MaIA documentation or talk to the team to design the flow with us.