ROAS or TACoS: what changes in e-commerce

ROAS and TACoS look like two versions of the same metric and are not. One looks inside the campaign; the other looks at the whole business. And it is entirely possible to have an excellent ROAS with a TACoS that worsens every month — the combination that fools the most e-commerce operations.

The two formulas

ROAS = revenue attributed to the campaign ÷ campaign spend.

A campaign with US$ 5,000 of spend and US$ 20,000 of attributed revenue has a ROAS of 4.

TACoS = total ad spend ÷ total revenue for the period.

If you spent US$ 5,000 and the store billed US$ 50,000 that month — counting ad sales, organic search, email, referral and repeat customers — TACoS is 10%.

The difference is the denominator. ROAS only knows what the platform attributed. TACoS knows everything that entered the till.

What each one answers

No row repeats. They are different questions, and using one metric to answer the other's question is the original mistake.

Why a good ROAS can hide a problem

Three common situations:

The campaign takes credit for sales that would have happened anyway. Remarketing to people already buying and brand search inflate ROAS without creating new revenue. The mechanism is in last-click attribution.

ROAS is high because spend is small. A campaign with little money picks off the easiest audience. Scaling drops the ROAS — and that is normal, not a failure. The criteria are in when to scale a campaign.

ROAS ignores margin. A ROAS of 4 on a product with a 20% margin does not pay the operation. The real math is in contribution margin and the metric itself in ROAS.

What TACoS tells you

TACoS is a trend metric, not a single-month one. What it says shows up in the direction:

TACoS falling with revenue rising. The best case. The brand is winning sales that do not depend on media: repeat purchases, direct search, referrals. Media is building something, not only buying transactions.

TACoS flat with revenue rising. Healthy, proportional growth. You are buying growth at a constant price.

TACoS rising with revenue flat. The warning sign. It costs more advertising to hold the same revenue, which usually means audience saturation or more aggressive competition.

TACoS rising with revenue rising. Depends on the phase. During a launch or entry into a new category, expected. In a mature operation, efficiency erosion.

A numeric example that settles it

Three months of the same store, with the same campaign running:

ROAS is identical across all three months. By the campaign's ruler, nothing changed and scaling is working.

Except TACoS rose four and a half points. Translated: the operation is increasingly dependent on advertising. Sales that do not come from media — repeat purchases, direct search, referrals — did not keep pace. The store is buying revenue instead of building a brand.

That may be exactly the chosen strategy, in which case nothing is wrong. The problem is reaching December without having noticed, because the only number on screen was a steady 4.5.

Where TACoS does not help

Worth being honest about the limit, because it is a large one.

TACoS does not point at a campaign. It is a whole-operation number and does not tell you where to act. Anyone trying to optimize a campaign by TACoS cannot, because the metric does not descend to that level.

It is also sensitive to seasonality. A Black Friday month changes the denominator in a way that does not compare to an ordinary month. The correct reading is always against the same month last year, or as a rolling average.

And it depends on total revenue being available. If you do not have the whole store's number, TACoS does not exist — which is why it is an owner's metric and not one for someone who only sees the ad account.

Using both in the same routine

1. Weekly, by campaign: ROAS and cost per result. The decision ruler inside media.
2. Monthly, for the operation: TACoS. One row per month, twelve months on screen. The trend is the information.
3. Monthly, the bridge: attributed revenue over total revenue. It shows how much of your revenue the platform is claiming, and usually reveals over-attribution.
4. Quarterly: margin. Because both ROAS and TACoS are meaningless if nobody knows what each sale leaves behind.

The combined reading with other profitability metrics is in marketing ROI.

The third number worth tracking

There is a companion metric that makes the pair much easier to read: attributed revenue as a share of total revenue.

It is the simplest possible calculation — revenue the platforms claim, divided by what the store actually billed — and it exposes something both ROAS and TACoS hide.

When platforms claim 40% of your revenue, that is a plausible picture. When Meta and Google together claim 110% of it, which happens more often than anyone admits, you are looking at double counting: both platforms credited the same sale, each under its own window and its own rules.

That number matters for two reasons. It tells you how much to discount the ROAS you are reading, and it tells you when a channel comparison has become meaningless. Two platforms each claiming the same customer cannot be compared on attributed revenue, no matter how carefully you build the table.

The fix is not a better attribution model. It is a source field filled at signup or checkout, and a monthly table of sales by source that comes from your own system rather than from anyone's dashboard. The standard is in UTM parameters for ads.

The summary

Use ROAS to decide inside the campaign and TACoS to decide about the operation. One answers "do I scale or pause this campaign"; the other answers "is my business getting less dependent on advertising, or more".

Watching only ROAS, you can spend a year with great campaigns and an operation that only grows while ads run. Watching only TACoS, you cannot act, because the metric points at nothing. Together they close the reading — and the document that puts them side by side is on the paid media reporting page.