ROAS vs ROI

People say “return” in ads meetings and mean three different ratios. ROAS and ROI are the ones that get swapped most often. They are not interchangeable. Using the wrong one is how a store scales ads that look efficient and still lose money after COGS.

If you need the ROAS formula and break-even logic first, read What Is ROAS?. This article is the comparison.

The two formulas

ROAS = ad-attributed revenue / ad spend

ROI = (gain − cost) / cost

For a simple ad campaign, operators often set:

• gain = attributed revenue (or contribution, if they are careful)
• cost = ad spend (or ad spend + variable fulfillment, if they are more careful)

Those two parentheticals are the whole argument. Compute ROAS with the ROAS calculator and ROI with the ROI calculator. Keep the inputs written next to the number.

Side-by-side

	ROAS	ROI
Question	How much revenue did we attribute per ad dollar?	After the costs we included, what did we earn per dollar invested?
Typical numerator	Platform or blended ad revenue	Profit or contribution (if defined)
Typical denominator	Media spend	Media spend, or a fuller cost pile
Good for	Bid strategy, creative tests, channel pacing	Go / no-go on a channel, inventory bets, “should we keep doing this?”
Blind spot	COGS, shipping, returns, incrementality	Slow; allocation fights; still inherits bad revenue if you feed it GMV
“4” means	$4 revenue per $1 ads (before costs)	400% return on the cost you defined—or 4× if you display it that way; agree the format

A campaign with 5× ROAS and 20% contribution margin has a simple modeled ad profit of:

Contribution from attributed sales − ad spend
= (0.20 × revenue) − spend

If revenue = 5 × spend, contribution = spend, profit ≈ 0. That 5× ROAS was break-even, not a win. The profit margin calculator is the missing piece in many ROAS screenshots.

When ROAS is the right question

Use ROAS when:

• You are allocating budget inside Meta, Google, or another network.
• Creative A vs B should be compared on the same attributed revenue definition.
• You have already set a break-even ROAS from contribution margin and you are pacing to it.

ROAS is a terrible answer to “are we a healthy business?” because it ignores everything that is not in the ads UI.

When ROI is the right question

Use ROI (or contribution after ads, which is often clearer than a percentage) when:

• You are deciding whether to keep a channel.
• You are comparing ads to a non-ad investment (3PL, photography, a developer sprint).
• Finance asks whether cash increased after all variable costs you agreed to include.

Define the cost list in writing: product, inbound, payments, outbound, estimated returns, marketplace fees, agency. If the list changes, the ROI changes. That is a feature.

Related metrics that settle arguments

• MER (marketing efficiency ratio): total store revenue / total ad spend. Blunt, harder to game with last-click, still not profit.
• CAC: spend to acquire a customer—see What Is Customer Acquisition Cost?. Better when AOV mix shifts.
• Contribution after ads: the number that should match how you think about survival.

Put the small set on one dashboard: E-Commerce Metrics That Matter. The e-commerce analytics guide is the longer map.

A working rule for meetings

Say the formula before the number. “ROAS 6 on 7-day click, gross sales, not returned” is a sentence you can debug. “ROI is great” is not.

If someone reports ROAS, ask two follow-ups: what revenue and what is break-even given margin. If someone reports ROI, ask which costs. That habit is worth more than another dashboard widget.

Worked comparison (illustrative)

Campaign spend: $2,000. Platform-attributed revenue: $8,000. ROAS = 4.

Landed COGS and variable selling costs on those orders: $5,600. Contribution before ads: $2,400. After ads: $400. If ROI is defined as that $400 / $2,000, ROI = 20%—fine, but not “4×.” If someone used GMV − ads only ($6,000 / $2,000), they would report 300% ROI and ignore COGS. Same campaign, three meeting-ready numbers. Only the middle one used contribution.

If half the $8,000 would have happened from branded search anyway, even the 20% is too high. Incrementality is not in the ROAS formula; you still have to ask.

Brand vs prospecting

Prospecting ROAS is usually worse than brand ROAS. That does not mean prospecting has worse ROI if brand is leftover demand. Allocate costs honestly: brand campaigns often harvest. ROI on “keep the brand campaign” might be high even when you should still cap it. ROI on prospecting might be low in month one and acceptable if you trust LTV—say so, with a horizon.

Do not average brand and prospecting ROAS into one KPI and then cut the only campaign that brings new customers.

When finance says ROI and marketing says ROAS, write both formulas on the same slide for one campaign, once. After that, the vocabulary fight usually ends. Use the profit margin calculator so the contribution input is not a guess.

Agency retainers belong in ROI’s cost list if they exist only to run ads. Leaving them out makes ROI look better than the P&L.