Blog

/

Products

Blog

/

/

Amazon DSP for Beauty and Personal Care: Repeat Math

Products

Use cases

Amazon DSP for Beauty and Personal Care: Repeat Math

Neha Bhuchar

Amazon DSP for beauty

Quick answer: Beauty and personal care is a good fit for Amazon DSP because the products get used up and bought again, but only if you pay for new customers with repeat purchases in mind. Before you spend, work out the most you can afford to pay for one new customer: first order contribution plus the repeat contribution you realistically expect. Then build three kinds of flights (launch, hero and replenishment), track every shade and size as its own ASIN, and check every product claim against Amazon’s creative rules before an ad goes live. Amazon DSP is now part of Amazon Ads Agent, with its controls under DVA+ advanced settings, but the strategy below applies to whichever name you use.

Key Takeaways

  • Repeat purchases are the case for DSP in beauty. The ad often loses money on the first order. Whether it pays back depends on how many new customers come back.

  • A worked example shows the gap. At a $22.22 cost per new customer, a brand with a weak repeat rate cannot afford the flight, while the same brand with a strong repeat rate can.

  • Replenishment audiences are your strongest lever. Time reminders to when the product runs out, and suppress recent buyers so you do not pay to remind people who just bought.

  • Amazon’s creative rules matter more in beauty. Claims need support, superlatives need a cited source and date, and direct comparisons with competitor brands are not allowed.

  • Variants break tracking. Every shade, size and pack is a separate ASIN. Sales of untracked variants can only appear as brand halo, and only when they share a subcategory with a tracked ASIN.

Why Beauty Brands Are Looking at Amazon DSP

Search ads in beauty have become more expensive. Beauty Independent reported on a study by the Amazon agency Envision Horizons that estimated branded search cost per click rose 248%, from $0.84 to $2.91, after large prestige brands expanded their spending, with non branded costs up about 186% over the past year. These are an agency’s own estimates, reported without independent verification, so treat them as a signal and not a benchmark.

The pressure is real for many beauty advertisers, and it pushes them to look at upper funnel and remarketing tools. But DSP does not become cheaper just because search is expensive. It has different economics, and it rewards brands that know their numbers. Our Amazon DSP guide covers how the platform works, and the rest of this article is specific to beauty and personal care.

What Changed with the DVA+ Rename

Amazon merged the Ads console and Amazon DSP into Amazon Ads Agent on September 29, 2026, and moved programmatic buying under DVA+, with the older DSP controls in advanced settings. Amazon says existing campaigns, audiences and performance history carry forward. DVA+ is still in closed beta in the US, so check what your account can use. Our guide to DVA, DVA+ and Full Funnel Campaigns covers the new names and what is still unconfirmed.

One point is specific to beauty. DVA+ builds AI generated creative for you to review. In a category with strict claim rules, treat that review as a compliance step, not a formality. The creative rules section below shows what to look for.

How Much Can You Pay for a New Beauty Customer?

Most beauty DSP flights fail or succeed on one number: the most you can afford to pay for a new customer.

The formula:

First order contribution = Order value × Contribution margin before ads


Expected repeat contribution = Share who reorder × Average reorders per reorderer × Contribution per order


Maximum affordable cost per new customer = First order contribution + Expected repeat contribution (within your payback window)


Worked example (illustrative). A serum sells for $28 with a 30% contribution margin before ads, so each order earns $8.40. You spend $4,000 on a prospecting flight and Amazon reports 180 new to brand customers, so your cost per new customer is $4,000 ÷ 180 = $22.22.

  • Weak repeat profile: 35% of new customers reorder, and each reorders 2.5 times in 12 months. Expected repeat contribution = 0.35 × 2.5 × $8.40 = $7.35. Maximum affordable cost = $8.40 + $7.35 = $15.75. You are paying $22.22, so the flight loses money even after repeats.

  • Strong repeat profile: 60% reorder, and each reorders 3 times. Expected repeat contribution = 0.60 × 3 × $8.40 = $15.12. Maximum affordable cost = $8.40 + $15.12 = $23.52. You are paying $22.22, so the flight pays back inside 12 months.


same flight falls

Figure 1: The same flight fails or passes depending on repeat behavior.

The numbers are illustrative, but the lesson is general. A product with good repeat behavior can justify a first order loss. A one time purchase cannot.

Where to get your own repeat data. Use your own order history by customer, not Amazon’s estimates. Amazon’s Long Term Sales metric is an estimate of incremental sales over the next 12 months, based on historical returns for customer engagements, and Amazon defines new to brand as no brand purchase in the previous 12 months. That makes it a useful reference but partly a forecast. Judge booked first order results first, and use your own repeat rate in the formula above.

Launch, Hero and Replenishment: Three Flights for Beauty

Beauty brands usually need three different flights. They have different goals, so they should not share a budget or a target.


three flights for beauty brands

Figure 2: Three flight types, three goals.

Flight

Goal

Audience

How to judge it

Launch

Awareness for a new item

Broad and interest based audiences, plus shoppers of related products

Reach, detail page views and new to brand share, not ROAS

Hero

Win new customers for your best seller

Prospecting audiences, lookalikes of past buyers, competitor and category shoppers

Cost per new customer against the maximum affordable cost above

Replenishment

Bring buyers back as the product runs out

Past purchasers, timed to usage

Repeat purchase rate and incremental sales through a holdout

Keep the budgets separate. A replenishment flight will often show the highest ROAS, but much of that credit goes to shoppers who would have reordered anyway. The hero flight will look worse on ROAS but is where the new customers come from.

Replenishment Audiences: Time the Reminder to the Product

Beauty products run out on a schedule. A 60 day supply of a serum, a 30 day supply of a cleanser and a six month sunscreen bottle all need different reminder windows.

Here is a starting rule you can test. It is our recommendation, not an Amazon setting:

Reminder window start = Days of supply minus about 20 days

Reminder window end = About 10 days after the supply runs out

Suppression window = The first 30 days after purchase

For a product with a 60 day supply, that means you remind past purchasers from around day 40 to day 70, and you suppress anyone who bought in the last 30 days.


Figure 3: A starting replenishment window for a 60 day product.

Then test it. Run one audience on the rule and one with a different start day, and compare repeat purchase rates with a holdout group that sees no ad. The winning window is the one your own data supports.

Two more controls matter in this category:

  • Frequency caps per audience. Past purchasers are a small group, so it is easy to show the same person the same ad too many times. Test caps rather than assuming one number.

  • Subscribe and Save. Amazon lists Subscribe and Save subscriptions among the on Amazon conversion events measured in attribution, so a flight can be judged on subscriptions as well as single orders, according to Amazon’s attribution documentation. A subscription is worth more than a single repeat, so track it separately.

Amazon’s Creative Rules for Beauty and Personal Care Claims

Claims are where beauty ads get rejected. Amazon’s creative acceptance policy has a health and beauty section and general rules that apply to the claims beauty brands love to make.

What the policy says

What it means for a beauty ad

Ads and landing pages cannot rely on claims that lack support or are overstated, and cannot set unrealistic expectations about effectiveness

Be careful with “visibly younger skin in 7 days” unless you can support it

Non prescription products cannot claim to be as effective as prescription products

Avoid “works like a prescription retinoid” on an over the counter product

Objective superlative claims require the source and date of the evidence to be cited

“Number one dermatologist recommended” needs a cited source and date

Comparative claims need an independent source dated within the previous 12 months, and direct comparisons to competitor brands are not permitted

Do not name a rival in the ad

Quantified performance claims must be verifiable on the product detail page

A figure in the ad must match the detail page

Ads cannot explicitly reference potentially embarrassing bodily functions or conditions

Wording for intimate care products needs extra care

Intimate personal healthcare products cannot run on certain placements, such as the Amazon homepage

Check placement limits before choosing inventory


Three notes. First, some of these rules are written for specific ad products, such as Sponsored Brands and display, so confirm which apply to your campaign type. Second, Amazon’s policy does not address before and after imagery, so check with Amazon Ads support before using it. Third, policies change, and your account manager is the source for the current version.

A pre flight claims check. Before any flight, list each claim in the creative and answer four questions: Is it a measurable claim or a feeling? Is the evidence on the detail page? Is the source and date cited where it should be? Does it name a competitor or imply a medical effect? If you cannot answer yes to the evidence question, change the claim.


five claim checks

Figure 4: The claims check before a flight goes live.

Track Every Shade, Size and Pack

Beauty catalogs are full of variants. A foundation has 30 shades, a shampoo has three sizes, and a gift set has its own listing. Each is a separate ASIN, and it matters for attribution.

When you build an order, you add the products you want to track. Those are your promoted products. Amazon can also credit brand halo, which it defines as products in the same brand and subcategory as a promoted ASIN. If you track only your hero shade, sales of other shades are only credited as halo, and only if they share a subcategory with the tracked ASIN. A gift set in a different subcategory may not be credited at all.

So make a tracked ASIN list part of your setup, and review it whenever you launch a shade or size. Our attribution guide explains promoted and halo metrics and the other rules that decide who gets credit.

Measuring Beauty DSP Without Fooling Yourself

Four checks keep the numbers honest:

  1. Judge prospecting on cost per new customer, against your maximum affordable cost, not on ROAS.

  2. Run a holdout on replenishment flights. They are the most likely to take credit for sales that would have happened.

  3. Wait for the data to settle. Amazon reports sales on the date of the ad interaction, and a report date is only truly settled about three weeks later, so recent days always look worse.

  4. Report new customers and repeat purchases separately from first order ROAS, and use your own order data for repeat rates.

Cost and Budget Considerations

Amazon has not published a fee change for DVA+, and says pricing carries forward, including self service and managed service options. Minimum spend varies, and managed service typically requires a minimum that differs by country. For a smaller beauty brand, that makes a test budget decision important: a flight needs enough spend to produce a reading, and you should set the pass or fail rule before it starts. Ask your account manager for current minimums.

How Atom11 Helps Beauty and Personal Care Brands

Atom11 is an Amazon Ads Partner platform built on the Amazon Ads API. Beauty DSP work is repetitive in a specific way: many products, many shades and sizes, several flight types running at once, and audiences that are small enough to over expose. Here is how Atom11 supports each part of the plan in this article.


Atom11 DSP


  • Separate orders for launch, hero and replenishment. The Amazon DSP module lets teams bulk create orders and ad groups, so you can set up each flight type by product line without building every order by hand. That matters when a launch adds a dozen shades at once.

  • Budget control across flights. Cross order pacing lets several orders share one monthly budget, so the hero flight can take spend while a weak launch flight slows down, without daily manual moves.

  • Frequency caps for small audiences. Past purchasers are a small group, so it is easy to show them the same ad too often. Frequency caps per audience let you test caps on replenishment audiences and keep them in place.

  • Audiences for replenishment and holdouts. AMC audience activation gives you a way to build past purchaser groups and the exposed and holdout groups your repeat purchase tests depend on.

  • DSP and search in one view. Side by side DSP and Sponsored Ads reporting, with promoted, halo and combined metrics, helps you see whether a hero flight is adding customers or simply taking credit from search, and it supports the halo checks for shades and sizes described above.

  • Stock awareness on the Sponsored Ads side. Retail aware rules use inventory, price, Buy Box and organic rank next to ad metrics, so Sponsored Ads spend can slow when a hero product runs short. That is useful for beauty brands whose best sellers can sell out quickly.

  • Safety nets and plain language analysis. Version control lets you save bids and budgets before a big event and roll them back afterward. The Atom11 MCP connects AI assistants such as Claude and ChatGPT to your ad accounts with user level permissions, live reads and writes that need confirmation, and setup takes about five minutes. Neo, Atom11’s AI analyst, compares periods across more than 15 ad and retail signals, which helps when you want to know why a flight moved.

Two honest limits. Amazon has said API access for DVA+ is targeted for Q1 2027, so confirm with the Atom11 team how DVA+ campaigns and existing DSP style orders are handled in the meantime. Also, your repeat purchase rate and replenishment timing come from your own order history, so no tool can supply them for you. The same questions apply to any platform you evaluate.

Atom11 DSP pricing starts at $499 a month for accounts spending under $50K and is 2% of ad spend above that, with the MCP and Neo included.

Want help setting up launch, hero and replenishment flights? Book an Atom11 demo.

Conclusion

Beauty and personal care suits Amazon DSP, but not for the reasons most guides give. The category works when you pay for new customers with repeat purchases in mind, so start with the maximum you can afford to pay for one, using your own repeat data. Then separate launch, hero and replenishment flights, time reminders to when the product runs out, and track every shade and size.

Keep claims on the right side of Amazon’s creative rules, especially now that DVA+ writes creative for you to review. And judge results on cost per new customer, holdouts and settled data, not on a single ROAS number. A beauty brand that knows its repeat rate before the flight starts will spend with more confidence than one that learns it from the dashboard afterward.

Find where your growth is leaking

See the same signal-driven automation against your own catalog and accounts.

Find where your growth is leaking

See the same signal-driven automation against your own catalog and accounts.

FAQs

Is Amazon DSP good for beauty and personal care brands?

It can be, especially for products people buy again. The case rests on repeat purchases, because the first order often loses money. Work out the most you can afford to pay for a new customer using your own repeat rate before you spend.

How do I calculate the maximum cost per new customer?

Add your first order contribution to the repeat contribution you expect inside your payback window. Repeat contribution is the share who reorder, times average reorders, times contribution per order. If your actual cost per new customer is above the total, the flight loses money.

What audiences should a beauty brand use on Amazon DSP?

Use broad and interest audiences for launches, prospecting and lookalike audiences for hero products, and past purchasers for replenishment. Time replenishment reminders to when the product runs out, and exclude recent buyers.

What are Amazon’s rules for beauty and personal care claims in ads?

Amazon’s creative acceptance policy says claims need support, objective superlatives need a cited source and date, comparative claims need an independent source from the previous 12 months, and direct competitor comparisons are not permitted. Confirm which rules apply to your campaign type.

Does Amazon allow before and after images in ads?

Amazon’s creative acceptance policy does not address before and after imagery in the sections we reviewed. Check with Amazon Ads support before using it, and keep any claims supported.

TABLE OF CONTENTS

Book a demo

See how Atom11 keeps ACoS in check

automatically.

Book a demo →

Enterprise Amazon advertising software for brands and agencies across the US, UK, EU and LATAM.

Copyright © 2026 Atom11. All rights reserved.

Enterprise Amazon advertising software for brands and agencies across the US, UK, EU and LATAM.

Copyright © 2026 Atom11. All rights reserved.

Enterprise Amazon advertising software for brands and agencies across the US, UK, EU and LATAM.

Copyright © 2026 Atom11. All rights reserved.