Advertising Cost of Sales (ACoS): How to Calculate It

How to calculate advertising cost of sales

The formula is simple: divide ad spend by the revenue those ads generated, then multiply by 100. As a made-up example, say you spend 2 million sum on Instagram ads in a month and the orders from them bring in 10 million sum. Your ACoS is 20%, meaning 20 of every 100 sum of revenue went on acquiring the customer.

The denominator must be revenue from advertising, not the revenue of the whole company. If you divide all marketing spend by total revenue, you get a blended figure. That is useful too, but it answers a different question: how heavy marketing is for the whole business, not how one campaign performed.

ACoS and ROAS are the same number flipped

ROAS shows how much revenue each sum of ad spend returns. ACoS shows how much of that revenue the ads consume. The relationship is ACoS = 1 / ROAS × 100. A ROAS of 4 means an ACoS of 25%, and a ROAS of 5 means 20%.

Advertisers tend to talk in ROAS, while owners often find ACoS easier because it can be compared with margin straight away.

What ACoS is safe for your business

There is no universal benchmark. Everything depends on margin. If a product earns a 30% margin and your ACoS is 35%, every ad-driven sale loses money. At 30% you break even on the margin but make no profit, because ads ate all of it. A workable rule: your acceptable ACoS must sit below the margin after other costs and the profit you want to keep.

Before comparing, take these out of the margin:

• delivery and packaging
• marketplace or payment fees, for example Click or Payme
• returns and cancelled orders
• discounts and promo codes
• salaries of managers and operators
• taxes
• the profit you want to keep

Track it by product and by channel

An average ACoS across the shop hides imbalances. A high-markup accessory can afford expensive clicks, while a fast-moving item with a thin margin may lose money at a modest ACoS. Calculate the metric for categories and key products. On marketplaces this helps you set bids for internal advertising: raise them on some products, lower them on others.

The same goes for channels. Instagram ads, Google search, Telegram broadcasts and OLX listings behave differently, so do not blend them into one figure. A channel with a high ACoS may bring new customers who buy again later, and it should be judged differently from a channel that attracts one-off bargain hunters.

Where ACoS can mislead you

The metric has limits worth knowing up front:

• It counts only the first purchase. If customers come back, their real value is higher, and the acceptable ACoS on a first order can be higher too.
• It depends on attribution. Someone sees an Instagram ad, messages you on Telegram and pays by phone a week later. Without end-to-end tracking the revenue goes to the wrong channel or disappears.
• It ignores brand effects. Awareness and organic orders do not show up in it.
• It does not include profit on its own. For that you need ROMI, which compares spend with profit rather than revenue.

Lowering ACoS without losing sales

You can push ACoS down from two sides: spend less to win each customer, or earn more from each order. In practice that means improving the conversion of your site or bot and the speed of manager replies, narrowing the audience, switching off weak ads, and raising the average order with bundles and add-ons.

The worst approach is to simply turn ads off to get a nice number. Sales may fall by more than you save, and total profit drops.

Make it part of your regular reporting

For active campaigns, check ACoS weekly. For the business as a whole, once a month is enough. A rising figure is a cue to check three things: whether competition in the auction has grown, whether bids have crept up, and whether conversion has dropped.

To keep the number honest you need data from several places: spend from ad accounts, orders from your CRM or shop, payments and offline sales. That is what end-to-end analytics brings together. If you want a report like this built around your business, the markdev team can help you work out which data you already have and what is missing.