What Amazon Marketing Cloud Actually Is (And Whether Your Brand Needs It)

Amazon Marketing Cloud (AMC) tells you what your customers are actually worth, and that one answer changes how you should spend on ads.

If you sell on Amazon, you've probably heard agencies and software companies talk about AMC as the next big thing. Most sellers I speak to still aren't sure what it does, whether they can use it, or whether it's worth the effort.

I've spent the last few months running AMC analyses across roughly ten consumable brands and fifteen hero products, mostly on Amazon UK. This guide is what I wish someone had given me at the start: what AMC is in plain English, what it shows that the ads console can't, and how to tell whether your brand needs it.

What AMC actually is

AMC is a secure data room where Amazon lets you ask questions of your own ad and purchase data at the customer level. You never see individual shoppers. You write a query, Amazon runs it, and you get back anonymised totals.

That sounds technical, so here's the difference in practice:

• The ads console tells you what happened to each campaign: clicks, spend, sales, ACoS.
• AMC tells you what happened to each customer: when they first bought, whether they came back, how often, what else they bought, and which ads they saw along the way.

A few things are worth knowing before you start:

• It's query-based. You ask questions in SQL. Amazon provides templates, and AI tools can now write most queries for you, but someone still needs to know which questions to ask.
• Results are aggregated. Amazon hides any result that covers too few customers. Small products need broader groupings, such as yearly rather than monthly cohorts.
• The best data sits in add-on datasets. Purchase history and Subscribe & Save data come from extra datasets such as Amazon Retail Purchases and Flexible Shopping Insights. Check what your instance has before planning any analysis.
• It's correlation, not proof. Attribution is last-touch. AMC shows you patterns, and big decisions still deserve a proper test.

What AMC shows that the ads console can't

The console judges every sale on the day it happens. AMC lets you judge the customer over their whole relationship with your brand. For consumable brands, that's where most of the money is.

These are the questions I now answer for almost every brand:

None of these can be answered from campaign reports alone. They all need data on individual customers over time, and that's what AMC holds.

What the data showed across ten consumable brands

These patterns came up again and again across supplements, personal care, hygiene and life-stage products. All figures are anonymised and shown as ranges.

1. The second order is where the value is

On every Seller Central consumable I analysed, a customer who came back for a second order was worth 2.8 to 4.5 times as much as a one-time buyer. Getting that repeat customer to subscribe to Subscribe & Save added only another 15% to 55%.

Most brands obsess over subscriptions. The bigger lever is almost always simply getting the second purchase.

2. The shape of the product predicts its lifetime value

Two products in the same account had lifetime values five times apart, because one was a daily habit and the other a one-off purchase.

If most of the value arrives in year 1, you don't need to wait years for the data to mature. If it keeps compounding, first-order ROAS is badly understating what a customer is worth.

3. Your ads are often re-buying customers you already have

Between 33% and 55% of ad-attributed purchases went to existing customers on several products. Branded search was consistently the worst:

On one product, branded search took half of all ad-attributed purchases and had the lowest new-customer rate in the account.

4. Customers reorder late

People reorder after they run out, not before. On a 30-day pack, non-subscribers came back roughly every 66 days, so they spent weeks without the product. Subscribers came back about every 41 days.

This tells you exactly when to show reorder ads: before the pack runs out, not after.

5. A high ACoS can still be profitable

For low-price, high-loyalty products, first-order ROAS understated the break-even ACoS by 1.5 to 2 times. Non-branded campaigns that looked unprofitable at a 45–49% ACoS paid for themselves within 8 to 9 months.

The account-level ACoS only looked disciplined because branded spend was holding it down.

6. Sometimes the right answer is to spend less

One product earned back less than 40p for every £1 spent acquiring a customer. No amount of optimisation fixes that. The analysis told the brand to stop scaling paid acquisition on that line, which saved far more than any bid change could.

Does your brand need AMC?

AMC pays off most for brands whose customers buy more than once. If yours mostly buy once, the console already tells you most of what you need.

AMC is likely worth it if:

• You sell a consumable or repeat-purchase product, such as supplements, personal care, pet food or cleaning products.
• You're spending enough on ads that a better budget split would matter.
• You're being held to a flat ACoS or first-order ROAS target that feels too tight.
• You use Subscribe & Save and aren't sure whether it's growing.
• You have several products and suspect customers buy across the range.
• You spend heavily on branded search and want to know how much of it is necessary.

It's probably not a priority if:

• Your products are bought once, such as furniture, one-off gadgets or gifts.
• Your sales volume is small. AMC hides results for small customer groups, so tiny products return little usable data.
• The basics of your account aren't in place yet. Fix the campaign structure and wasted spend first.

A quick test: if you don't know roughly what a customer is worth after 12 months, and you sell something people reorder, AMC will probably change how you spend.

How to get started, and the mistakes to avoid

Start with three questions, not thirty: what is a customer worth, how long until they come back, and what share of ad sales go to new customers. Those three alone usually change the budget conversation.

A simple order to work in:

1. Check your data first. Confirm which datasets your AMC instance has and how far back they go. List every product, including old listings that were replaced.
2. Run the core numbers. Lifetime value by customer group, time between orders, new-to-brand share by campaign type.
3. Look at subscriptions if you use Subscribe & Save: when people subscribe, and how many cancel.
4. Turn it into decisions. Budget split, retargeting timing, bundle and cross-sell ideas.
5. Run one query at a time, and understand each result before running the next.

Mistakes I've made or nearly made:

• Treating a price drop as a change in behaviour. One product's customer value fell about 40%, but units per customer barely moved. The cause was deeper discounting.
• Counting relisted products as new customers. Customers moving from an old listing to a new one showed up as brand-new buyers, making up to a quarter of early customer groups.
• Setting the wrong date window. Running a "first purchase" query over too short a window labels existing customers as new. In one case it inflated new customers by about 60%.
• Comparing Subscribe & Save across April 2026. Amazon stopped pre-selecting Subscribe & Save that month, and first-order sign-up rates fell by roughly a third on every product I measured. Comparisons that span the change are misleading.
• Mixing up two kinds of acquisition cost. Cost per order and cost per new customer can differ by 2 to 3 times on the same product.

The bottom line

AMC won't replace good campaign management, but it tells you what good looks like for your brand. When you know what a customer is worth and how long they take to come back, ACoS targets, budget caps and branded spend all start to look different.

Over the coming weeks I'll go deeper into each of these: calculating Amazon customer lifetime value, timing ads around reorder cycles, and what new-to-brand data really tells you.

Want to see what your customers are worth? Subscribe free to get new guides by email and unlock the Amazon LTV calculator.