โ† Back to Blogs

Amazon Ads

9 min read

Your ACoS Target Is Ignoring What a Repeat Customer Is Worth

Onieque Edwards

Content Strategist /Blog Writer

Most Amazon ad targets are set on one number: the ACoS at which the first sale breaks even.

For a product someone buys once, that is a sensible place to start. For a product someone uses up and buys again, it leaves out most of what the customer is worth.

This post walks through the arithmetic slowly, with one example product and every figure written down. All the numbers below are illustrative. They are round on purpose, they come from no real account, and your own will be different. The point is the method, so you can rebuild it with yours.

Why first-order ACoS reads a consumable wrong

ACoS is ad spend divided by the ad-attributed sales it produced. It tells you whether an ad paid for itself on the order it was credited with.

That is a fair test for a durable product. A shopper who buys a desk lamp may not buy from you again for years, so the first order is close to the whole value of the customer.

A supplement, a pet treat or a coffee pod works differently. The first order is where the relationship starts. The customer runs out, then reorders through Buy Again, Subscribe & Save, or a search for your brand name. Those later orders are usually not credited back to the ad that found the customer, so an ACoS report never sees them.

The ACoS number is correct. It just stops counting after the first order.

The $30 supplement, one line at a time

Here is the example product. A bottle of a daily supplement, one month's supply, sold for $30.

Take one order and subtract what it costs to fulfil it:

  • Price: $30.00

  • Referral fee: about $4.50, taken here as 15% of the price

  • FBA fee: $5.50, an example figure for a small, light bottle

  • Cost of goods: $7.00 to make the bottle and land it in the warehouse

  • Left before advertising: $13.00

That $13 is what a first-order target is built on. If you spend $13 in ads to win the sale, you break even on it. That is an ACoS of about 43%. Many brands set their cap below that so each sale shows a profit, at around 30% or 35%.

Now follow the customer for a year. Say the typical customer in this example places four orders in twelve months: the first one plus three reorders.

Say the three reorders come through Subscribe & Save at a 5% discount. That takes $1.50 off the price, so each reorder sells for $28.50. The referral fee falls a little with the price, to about $4.28. The FBA fee and the goods cost stay the same. Each reorder leaves about $11.72.

For the example, assume none of those reorders needed an ad. The customer simply reordered.

Add it up:

  • First order: $13.00

  • Three reorders at about $11.72: about $35.18

  • Twelve-month margin from one customer: about $48

So the first-order view prices this customer at $13. The twelve-month view prices them at about $48, close to four times as much. Every bid in the account is being set against the smaller figure.

Revenue value and profit value are different numbers

Customer value can be built in two ways, and they give very different answers.

Revenue-based value adds up what the customer spends. In the example that is $30 plus three reorders at $28.50, about $115.50 a year.

Profit-based value adds up what the customer leaves you after fees, goods and discounts. In the example that is about $48.

Both figures describe the same customer. But only one of them can pay for advertising.

The revenue figure is attractive because it is easy to get and it is large. It is also the version that gets brands into trouble. Suppose someone sets their acquisition ceiling at half of revenue value. That comes to about $57.75 to win one customer, which is more than the whole $48 the customer will ever leave behind. Every new customer won at that price loses money, and the loss only shows up months later.

So the first-order cap and the revenue-based value make opposite mistakes. The first prices the customer too low and loses good customers to whoever bids more. The second prices them too high and buys customers at a loss. Profit-based value is the only one of the three a bid can safely sit on.

Two things belong inside the margin figure that are easy to leave out.

The discounts you actually give. Subscribe & Save lets sellers fund a discount, typically 5% or 10%, and coupons and promotions can land on the same order. Since 23 July 2026 there is also Buy Again & Save, aimed at Prime shoppers who reorder without subscribing. Public reports differ on whether it adds a new seller-funded discount or reuses the funding you already set for Subscribe & Save. Check your own Subscribe & Save settings and Seller Central news before you assume what a reorder now earns.

Any ad cost on reorders. The example assumed the reorders were free. If some customers come back through a branded search ad or a retargeting campaign, that spend comes off the reorder margin.

Setting a safe maximum acquisition cost

The twelve-month margin is the most a new customer could be worth. It is not what you should pay for one.

Pay the full $48 and you only break even at the end of the year, and only if every assumption held. A safe maximum leaves room for the estimate to be wrong.

A working rule is to cap acquisition cost at half to two thirds of twelve-month margin. Go nearer half when cash is tight or when the repeat data is thin. In the example, that puts the safe maximum at about $24 to $32 to win one new customer.

Read that range against the first-order view. $24 on a $30 first order is an ACoS of 80%. $32 is above 100%. In a report that only knows first orders, both look like mistakes. Read against the year, $24 is about half of what the customer leaves you, and the business still keeps the other half.

One detail matters here. The safe maximum is a cost per new customer, not per ad order. Some ad orders come from people who already buy from you. Amazon Ads counts a shopper as new-to-brand when they have not bought from the brand in the last 12 months. So the honest test is ad spend divided by new-to-brand orders, measured against the ceiling. Dividing by all ad orders makes the cost of a new customer look cheaper than it is.

What makes the number unsafe

The safe maximum is only as good as the repeat figure under it. These are the usual ways that figure goes wrong.

The window is too short. A customer acquired three weeks ago has not had time to reorder a 30-day product. Judge repeat behaviour over at least twice the time one unit lasts. For a twelve-month value, use customers who were won at least a year ago, so the year has actually happened.

The repeat rate was assumed, not measured. If the brand is registered, the Repeat Purchase Behavior report in Brand Analytics shows orders and unique customers. Orders divided by unique customers, over a long window, gives a working orders-per-customer figure for your brand. Read brand-level and ASIN-level numbers separately, because a customer who switches to a different product of yours counts as a repeat brand customer but not a repeat ASIN customer.

Only subscribers were counted. The Subscribe & Save dashboard, upgraded in May 2026, shows average reorders per subscriber and splits reorder share from Subscribe & Save share of units. Subscribers are only part of the picture. One CPG brand, reported by Incrementum Digital through My Amazon Guy, had a 32% reorder rate while Subscribe & Save share sat in the low single digits. Build the value from subscriber data alone and it will come out too low.

Returns were not taken out. Repeat Purchase Behavior revenue does not net out returns. In a category where returns are common, adjust before you trust it.

Stock did not hold. A twelve-month value that assumes three reorders falls apart if the product is out of stock for one of them. The Subscribe & Save dashboard shows the share of scheduled units not delivered because of a stockout. If that share is meaningful, your repeat figure is already lower than the average suggests.

If one of these cannot be checked, use the low end of the range or hold the ceiling where it is. An acquisition cost built on a repeat rate nobody measured is a guess with a bid attached.

What to do with the money

The decision is where the next dollar of ad budget goes.

A brand that judges everything on first-order ACoS sends its money to the searches that look efficient today. Those are often branded searches and past buyers, many of whom were coming back anyway. It pulls money away from the category searches where new customers are found, because a first order there rarely breaks even on its own.

A brand that knows its twelve-month margin per customer can do the opposite on its proven reorder products. It can raise the ceiling on category searches up to the safe maximum, measured per new-to-brand order. It can hold branded and retargeting spend to a tight return, because that money is mostly buying reorders the brand would have kept.

If the repeat figure is not measured yet, that is the first money to spend: an hour with the Repeat Purchase Behavior report before a single bid moves. Jungle Pundits builds this number for consumable brands as part of an ads review, and it is often the line that changes the budget the most.

The related question is who wins the customer when your cap refuses to pay. That is the subject of our companion post, Your First-Order ACoS Cap Is Paying a Competitor To Keep Your Customer.

FAQ

Why does first-order ACoS undervalue consumable products?

Because it counts only the order the ad was credited with. A consumable customer reorders through Buy Again, Subscribe & Save or brand search, and those orders are usually not credited to the ad that found them. The value is real, but the ACoS report never sees it.

Should customer lifetime value be based on revenue or profit?

Profit. Revenue value adds up what the customer spends, which is far larger than what they leave you after fees, goods and discounts. In this post's example, a customer worth about $115.50 in revenue leaves about $48 in margin. A ceiling set on revenue can cost more than the customer will ever return.

How do I calculate twelve-month margin per customer on Amazon?

Work out what one order leaves after the referral fee, the FBA fee, the cost of goods and any discount you fund. Multiply by orders per customer over twelve months, using the lower margin for discounted reorders. Take off any ad cost spent on those reorders. If you are brand registered, the Repeat Purchase Behavior report gives orders and unique customers to estimate orders per customer.

What is a safe maximum acquisition cost?

A working rule is half to two thirds of twelve-month margin per customer, nearer half if cash is tight or the repeat data is thin. Measure it as ad spend per new-to-brand order, not per ad order.

Can an ACoS above 100% ever make sense?

On first orders from genuinely new customers, sometimes. In this post's example, a safe maximum of about $32 on a $30 first order is an ACoS above 100%, and it still leaves margin over the year. It only holds if the repeat rate is measured, stock holds, and the spend is winning new-to-brand customers rather than re-buying existing ones.

Sources

Amazon Ads defines new-to-brand shoppers as those who have not bought from a brand in the last 12 months, in its announcement of the long-term sales metric. Amazon's overview of Brand Analytics lists the Repeat Purchase Behavior report. Amazon's guide What is Amazon Subscribe & Save? covers the seller-funded discount and the Dashboard tab in Seller Central. My Amazon Guy reported the May 2026 Subscribe & Save dashboard update, including reorders per subscriber, reorder share against Subscribe & Save share, and the Incrementum Digital example, and the Buy Again & Save launch on 23 July 2026.

Onieque Edwards

Content Strategist /Blog Writer

Onieque is the brain behind bold Amazon growth strategies and structured business execution. He enjoys turning scattered ideas into clear, actionable systems that actually drive results. When heโ€™s not building out growth plans or refining campaigns, youโ€™ll likely find him exploring new coffee spots or getting lost in ideas that connect strategy with creativity.