Why beauty buyers stop at one
A shopper buys a foundation. The shade is slightly off. Or she buys a cleanser, uses it for a week, and is not sure it suits her skin. Then she gets a generic "thanks for your order" email and a 10% coupon for anything in the store.
She does not reorder. She returns the foundation, or leaves it in a drawer. Your Meta spend bought one order at a loss, and her next purchase goes to the brand whose ad happens to reach her next.
Beauty has two problems most categories do not: fit (shade, skin type, sensitivity) and routine (products used together, in order). A post-purchase system that ignores both cannot work. The one in our case study was built around them.
Shade and skin-type quiz data into Klaviyo
A quiz is only useful if its answers reach the email platform. We connect the quiz app to Klaviyo so every answer lands as a profile property:
| Quiz answer | Klaviyo property | Where it changes the email |
|---|---|---|
| Skin type (dry, oily, combination, normal) | skin_type | Welcome flow branch, product block filter |
| Main concern (dullness, texture, redness, dryness) | skin_concern | Which routine step is recommended next |
| Shade and undertone | shade, undertone | Shade-check email, complexion launches |
| Sensitivity | sensitive_skin | Excludes acids and fragrance from recommendations |
| Current routine length | routine_steps | How many products we suggest at once |
The welcome flow for a quiz taker who has not bought yet runs four emails over 10 days. Each one names her result: "your routine for combination skin," not "our bestsellers." For shoppers who skip the quiz, the second welcome email invites them to take it, because a profile with a skin type is worth more to every later flow.
Two rules we hold to. Never recommend something the customer said she cannot use. And keep shade data current: if she buys a different shade later, the property updates from the order, not from the old quiz answer.
The post-purchase sequence
This is the shape of a beauty post-purchase sequence as we build it. It starts at delivery, not at order, because nobody can judge a moisturizer that has not arrived.
| Timing | Job | |
|---|---|---|
| Delivery day | How to use it | Amount, order in routine, AM or PM. A short application video. |
| Day 3 after delivery | Shade or skin check | "Did the shade match?" with a one-click exchange link for complexion products. |
| Day 7 | What to expect | When most people notice a difference, and what is normal in week one. |
| Day 14 to 21 | Next routine step | One adjacent product, chosen by skin type and what she bought. |
| Day 21 | Review request | Asks for a photo review, filtered by skin type for later reuse. |
| Replenishment date | Running low | Timed to that product's real reorder curve (below). |
The shade-check email deserves attention. A shopper with the wrong shade will either return it or quietly stop buying. Offering the exchange first, before she has to ask, saves the order and tells the brand which shade descriptions mislead people.
Regimen cross-sell timing
Cross-selling in skincare is about the routine, not the catalog. The next product should be the step that sits next to what she already uses:
- Cleanser buyers get a serum or treatment matched to their concern.
- Serum buyers get the moisturizer that layers with it.
- Moisturizer buyers get the cleanser, because a routine with a gap at the start is easy to abandon.
- Makeup buyers get the matching primer, setting product or remover, keyed to shade.
Timing matters as much as product. Cross-sell too early and the customer has no opinion of the first product yet. Too late and she has already settled into another brand for the missing step. Day 14 to 21 after delivery is where we start, then adjust from the brand's own order data.
We also check for conflicts. A customer who flagged sensitive skin and bought a gentle cleanser should not be pushed toward a strong exfoliating acid in the next email.
Replenishment by product
A 30ml serum and a 200ml cleanser do not run out on the same day, and two people use the same cream at different rates. So we do not guess. We pull from Shopify the median number of days between repeat orders of each SKU and set the first reminder about a week before that median.
For a new product with no repeat data, we estimate from the size and the directions ("two pumps morning and night") and replace the estimate once 50 or so repeat orders exist. Each reminder offers the same size, a larger size where one exists, and, for proven repeaters, a subscription.
Sampling that converts
Beauty brands give away a lot of product. Most of it is never followed up.
- Tag every sample at fulfillment. Whether it is a deluxe sample in an order or a paid trial kit, the shipped sample is written to the order so Klaviyo knows who has what.
- Follow up on that exact product, about a week after delivery: how to use it, how long the sample lasts, and the full-size link.
- Credit trial kits toward full size. A paid kit that credits its price against the first full-size order gives the customer a reason to come back.
- Report sample-to-full-size conversion per product. Some samples convert. Others are just a cost, and those should stop.
UGC creative testing on Meta
The case study's 34% CPM reduction came from UGC-led creative. Our testing works like this:
- Concepts first. Each week we test new concepts, each one a different reason to buy: texture, wear time, a skin-type story, a routine, a comparison with what she uses now.
- Hooks second. Winning concepts get three to five new opening seconds. In beauty the first frame is usually the product on skin, in natural light.
- Creators across skin tones and types. A foundation ad shown only on one skin tone sells one shade. We brief creators to match the shade range we need to move.
- Winners move to scaling campaigns, and we retire them as frequency climbs and results fade.
Angles that work in beauty and rarely in other categories: the swatch-on-arm shot, the "wear test" at hour one and hour eight, the "what is in my bag" format, and an esthetician or makeup artist explaining an ingredient list in plain words.
Influencer whitelisting and partnership ads
Whitelisting means running your paid ads from a creator's account, with their permission. The ad carries their name and face, so it reads as a recommendation rather than a brand ad. Meta runs this through partnership ads.
We use it for creators whose organic posts already drive sales. The creator grants access, we run their best posts as ads with our targeting and budget, and we pay them for usage on agreed terms and dates.
Disclosure still applies. The FTC's Endorsement Guides FAQ says a free product or payment is a material connection that must be disclosed, that a platform's built-in tool is not automatically enough, and that the brand is ultimately responsible for what others do on its behalf. We check every whitelisted ad for a clear disclosure before it runs.
Cosmetic or drug: watch the claims
In the US, whether a product is a cosmetic or a drug depends on its intended use, and intended use is shown by what you claim in ads and email, not only on the label. The FDA's guide, Is It a Cosmetic, a Drug, or Both?, lists claims such as restoring hair growth or reducing cellulite as drug claims, and treats acne cleansers and moisturizers with sun-protection claims as drugs or cosmetic and drug products.
So "clears acne" in a subject line is not a copy choice. It changes how the product is regulated. Our writers work from an approved claim list, and the post-purchase "what to expect" email describes how skin looks and feels, never what condition it treats.
Product page CRO for beauty
- Swatches on several skin tones, photographed in the same light, beside the shade selector.
- A shade finder or quiz link on every complexion product page, not only in the nav.
- Reviews filterable by skin type and shade, using the same properties the quiz collects.
- Full ingredient list as text, plus a short "free from" line for shoppers who scan for fragrance or specific actives.
- Routine bundles on the page of each step, priced below the sum of the parts.
Sorted has worked with 100+ D2C brands that did $425M in Shopify revenue in 2025, with $30M+ in managed ad spend at a 4.5x average ROAS. Agencies can have us deliver this beauty playbook under their brand: see white-label for agencies. Other categories are on the industries page.