ROAS vs ROI: What Your Ad Revenue Leaves After Costs

ROAS tells you how much revenue is attributed to each dollar of advertising spend. ROI tells you how much remains after the costs included in your calculation, compared with those costs. A campaign can have an impressive ROAS and still lose money.

This guide covers ROAS and ROI for advertising spend. To estimate what a customer is worth over the whole relationship, see how to estimate customer lifetime value and contribution.

The quickest way to understand the difference is to follow the same money through both calculations.

The phrase “included costs” matters. An ROI calculation covering advertising alone answers a narrower question than one covering advertising, campaign services and the cost of supplying the orders. Ask to see the formula whenever someone presents an ROI percentage.

One campaign, 5× ROAS and $1,800 left

Consider a fictional Canadian retailer. All amounts below are Canadian dollars, excluding sales tax. These are teaching figures, not client results, typical costs or recommended targets.

The retailer spends $6,000 on advertising. Its campaign is credited with $30,000 in orders after checkout discounts but before refunds. That gives it a 5× ROAS: $30,000 divided by $6,000.

Now follow what happens to that revenue.

The campaign has two revenue ROAS figures, depending on whether refunds have been deducted:

• Before refunds: $30,000 ÷ $6,000 = 5× ROAS.
• After refunds: $27,000 ÷ $6,000 = 4.5× ROAS.

For the ROI calculation, the included costs total $25,200: $16,200 in fulfilment and $9,000 in marketing. The return is therefore $1,800 ÷ $25,200 = 7.14% ROI on the costs included.

Both calculations can be correct. They describe different parts of the same result. The retailer retained $4.50 in revenue per advertising dollar, then used most of that revenue to cover fulfilment and marketing.

The $1,800 still has to contribute towards any company costs excluded from this example, including unallocated overhead, financing and income tax. It should be described as contribution after marketing, rather than company-wide net profit.

Try the example in the workbook

Download the ROAS and ROI reconciliation workbook. The supplied example contains the figures above.

Open the Bridge tab to see revenue, costs and the resulting ratios together. To use your own figures, save a copy, read ReadMe, replace the editable entries in Inputs and CostRegister, and complete the checks that apply to your records. Leave formula cells unchanged. Detailed help with the workbook appears later in this guide.

The workbook has been checked in LibreOffice. Microsoft Excel desktop and web have not been tested. Enter literal values into input cells and paste values only; rejection of formulas or TRUE/FALSE values pasted into inputs is not guaranteed. Review the result in your spreadsheet application before relying on it.

Why “500% ROAS” does not mean “500% profit”

ROAS is sometimes displayed as a multiple and sometimes as a percentage. A 5× ROAS and a 500% ROAS describe the same revenue-to-ad-spend ratio. Each advertising dollar is associated with five revenue dollars.

Those five dollars include the money needed to recover the advertising dollar itself. They also have to cover the product or service sold, plus whichever other costs apply. Treating the whole five dollars as a return after costs overstates what is available.

Even subtracting advertising alone leaves an incomplete answer for our retailer. Taking $30,000 of orders and subtracting $6,000 of media produces $24,000. That calculation has omitted refunds, fulfilment and campaign support. Calling the $24,000 profit would erase nearly all the economic work needed to generate those orders.

ROI starts after the relevant costs have been deducted. A 7.14% return on $25,200 of included costs means approximately seven cents remain for every cost dollar in that calculation. The business recovers the cost dollar and has that additional amount left within the stated scope.

There is another percentage that often causes confusion: margin. The retailer’s $1,800 contribution after marketing is 6.67% of its $27,000 retained revenue. That is a contribution margin after marketing. Its ROI is 7.14% because ROI divides by the included costs instead of revenue. The underlying dollars have not changed.

You do not need to put every percentage on a monthly report. For most readers, retained revenue, total included costs, contribution after marketing, ROAS and one clearly defined ROI are enough. Additional ratios become useful when they answer a specific question, such as comparing marketing efficiency or explaining product mix.

Google’s ROI explanation illustrates including production and advertising costs. It is a simplified advertising example, so a business still needs to define its own cost scope. A clear equation is more useful than assuming everyone means the same thing by ROI.

Why another report might show 20% ROI

Suppose the retailer’s marketing team divides the same $1,800 contribution by its $9,000 marketing investment. The answer is 20%.

That calculation measures contribution after marketing relative to marketing cost. The workbook calls it “marketing-cost contribution return” to keep the denominator visible. The 7.14% figure includes fulfilment costs in its denominator; the 20% figure does not.

Neither percentage becomes more truthful because it is larger or smaller. They must retain their definitions when compared, especially across agencies, channels or reporting periods. If a report switches denominators, show that change explicitly. Otherwise a reported improvement could be a change in arithmetic rather than a change in the business.

Which number should you use?

Use revenue ROAS to examine the relationship between attributed sales value and media spend. It is helpful when comparing advertising activity that uses the same conversion definition, revenue treatment, reporting period and attribution method.

Use a cost-complete return calculation when the question is whether the resulting business leaves enough money after the relevant costs. That needs information outside most advertising dashboards: fulfilment costs, campaign support costs, refunds and any other costs included in the decision.

For our retailer, a marketing manager checking day-to-day delivery might watch revenue ROAS alongside conversion volume. A finance discussion about the campaign’s contribution needs the $27,000 revenue and $25,200 cost calculation. A discussion about increasing the budget needs additional evidence about what another dollar of spending is likely to produce.

Those are different decisions. Making every report use a single percentage usually removes useful information.

The reporting level matters too. A small ad group may have too few completed purchases to support a stable comparison, even if its ratio can be calculated. One unusually large order can dominate its result. At that level, look at the underlying order count and values before describing a higher ROAS as a reliable pattern.

A broader campaign report can make the economics easier to see, but it can also hide meaningful differences. A campaign selling both low-margin and high-margin products might show an acceptable average while one part consumes contribution. Keep the overall result and investigate the product or service differences that could actually change a decision.

A practical reporting choice is to put the monetary outcome first: “$27,000 retained revenue, $1,800 contribution after the included costs.” Then add the ROAS and ROI as summaries. That gives readers enough context to understand a percentage without having to reverse-engineer it.

A higher ROAS can come with less contribution

Suppose the same retailer reduces media spending to $3,000 in a separate scenario. Retained revenue falls to $15,000, while the contribution margin before marketing remains 40%. Assume its $3,000 of management, creative and measurement costs remain unchanged for the period.

Revenue ROAS improves to 5×: $15,000 divided by $3,000. Yet the retailer now has only $6,000 before marketing. Advertising and the other marketing costs consume that whole amount, leaving zero contribution after marketing. The base case’s lower 4.5× net ROAS left $1,800.

That comparison does not prove the retailer should restore its old budget. It demonstrates why the ratio and the dollars should be read together. A campaign can become more efficient relative to media spend while producing less contribution overall, especially where some supporting costs remain fixed.

It also explains why a target ROAS is not a complete business objective. A target describes a desired value-to-media-cost relationship. The business still cares about the amount of useful revenue, the costs of earning it and the capacity to serve it. A report that celebrates the percentage while omitting those amounts can reward the wrong trade-off.

When reviewing a change, put the old and new revenue, media costs, other costs and contribution side by side. Ask which amounts changed before judging the ratio. If the higher ROAS came from withdrawing from lower-return sales, determine whether those sales nevertheless contributed towards fixed costs. If the lower spend also allowed a support cost to fall, include that saving rather than assuming the old cost must remain.

The key comparison is between complete alternatives under their actual cost conditions. Holding a cost fixed in an example helps isolate the mechanism; a real decision needs evidence about whether that cost really stays fixed.

Start with the revenue your business actually retains

The revenue number deserves as much attention as the formula. Before calculating either measure, check what the advertising report calls a conversion and what value it assigns to that conversion.

In Google’s Target ROAS documentation, the bidding objective is expressed through conversion value relative to cost. That value could reflect purchase revenue or another value supplied to the system. The label alone cannot tell you whether the number represents completed sales, estimated lead value or a profit-based value.

A retailer might pass purchase revenue. A service business might assign a value to a qualified enquiry. Another advertiser might supply a contribution estimate. All three can produce a value-to-cost ratio, but putting them in one column labelled “sales ROAS” would conceal important differences.

For a revenue calculation, work from a clear sales definition. Our example includes accepted order value after discounts and before refunds, then deducts the refunds separately. It excludes sales tax and contains no value for leads or predicted repeat purchases.

If an order export already subtracts refunds, do not subtract them again. If it records the original full price, adjust for discounts before treating that number as the amount customers agreed to pay. Where delivery charges are included in revenue, make sure the cost side includes the relevant delivery expense on a consistent basis.

Currency is equally basic and easy to miss. A dollar sign can mean CAD or USD. Convert source amounts under one documented policy before combining them. An apparent change in campaign return can otherwise be partly a currency movement or simply a mixed-currency error.

Tax treatment should be consistent with the business’s accounting approach. The examples here exclude sales tax and make no assumption about your ability to recover tax on purchases. Ask the person responsible for your accounts how revenue and costs should be presented together.

Keep future sales out of today’s completed-sales figure

A repeat customer can make an acquisition worthwhile over a longer period. That is a reason to analyze customer value, not permission to add an unsupported lifetime estimate to completed order revenue.

If the retailer expects further purchases from its new customers, show those as a separate forecast with a defined time horizon. Include the cost of supplying and retaining those customers. A future revenue estimate without future costs can make a weak acquisition look far more attractive than it is.

When actual repeat purchases arrive, replace or reconcile the relevant estimate. Adding the realized sales on top of the original forecast for the same purchases counts the expected benefit twice.

Use the customer lifetime value calculator when that longer-horizon question matters. Keep the current campaign reconciliation understandable on its own: observed sales in one view, future customer-value assumptions in another.

Put the missing costs back into the picture

The retailer spends $6,000 buying advertising, but marketing costs total $9,000. The additional $3,000 consists of $1,500 for management, $1,000 for creative and $500 for measurement in this fictional example.

Those costs do not belong inside media spend if you want an ordinary revenue ROAS calculation. They do belong in the broader return calculation when they are part of the campaign’s stated cost scope. Keeping both totals visible prevents a management invoice from disappearing between the dashboard and the business report.

Fulfilment is a separate group. Depending on the business, the relevant variable costs may include stock, packaging, payment processing, delivery subsidies, commissions or labour that changes with the work supplied. BDC defines variable costs by how they change with goods produced or services provided. The actual behaviour of your costs determines the classification.

A fixed monthly software fee, for example, does not necessarily increase because the campaign produces one more sale. A transaction fee does. A staff member’s salary may remain the same across a range of order volumes, while overtime or an additional hire changes the cost once capacity is exceeded.

This matters when you move from describing last month’s result to considering more spending. An allocated share of existing resources can be appropriate for understanding the resources a campaign uses. A decision about an additional campaign also needs to identify which costs would actually increase or become avoidable.

Shared costs need a fair allocation

If one creative project supports several campaigns, charging the full invoice to each campaign inflates total cost. Omitting it from every campaign understates the resources used. Choose an allocation that reflects how the work is used and ensure the allocated shares add up to the original cost.

For example, a campaign-specific landing page could be assigned directly to that campaign. Shared management work might be allocated using recorded time. Reusable creative might need a defined allocation across its expected use. These are possible methods, rather than a universal accounting rule.

Use the method consistently. A change in allocation can alter campaign-level ROI without changing total business spending. If a campaign looks worse only because it received a larger share of a shared expense this month, explain that before concluding its ads deteriorated.

There is no need to invent an elaborate allocation for every small cost. Concentrate first on amounts that could change the decision, while stating what the result excludes. Have a finance lead review material classifications and shared-cost choices before the calculation supports a spending commitment.

The word “total” in the workbook means all costs included in this model. It does not imply that every company expense has been allocated or that the remaining contribution equals the profit on the company’s financial statements.

How a 4× ROAS can still lose money

Return to the same fictional retailer. Imagine refunds rise from $3,000 to $6,000, while the net fulfilment cost remains $16,200 and marketing still costs $9,000.

The retailer now retains $24,000 of its $30,000 in initial orders. Its revenue ROAS after refunds is still 4×: $24,000 divided by $6,000 of advertising. That can look reassuring when read in isolation.

The full calculation tells a different story:

• Retained revenue: $24,000.
• Included fulfilment and marketing costs: $25,200.
• Contribution after marketing: −$1,200.
• ROI on included costs: −$1,200 ÷ $25,200 = −4.76%.

The loss arises because this scenario assumes the retailer cannot recover additional fulfilment costs when those extra orders are refunded. Perhaps goods cannot be resold, outbound shipping is unrecoverable or processing costs remain. The illustration does not assume this happens to every returned order.

A return has at least two effects to investigate. The refund reduces revenue. The returned product or cancelled service can also change costs. Recovered stock may reduce the net product cost, while return shipping or handling may add an expense. Those cost adjustments belong on their own lines.

Deducting the refund from revenue and then adding the same refund as an expense would count it twice. Reducing every cost automatically by the refund percentage can create the opposite problem, assuming recoveries that never happened.

Google provides reporting for the impact of conversion adjustments. That capability does not establish that your own order changes have been sent to the advertising account correctly. Check the connection between refunded orders, the sales records and the platform’s current conversion values.

The useful next question is specific: how much revenue was retained, and which costs were recovered or added? A general instruction to “improve ROAS” would not diagnose this example. The pressure on the result sits in returns and their cost consequences.

The initial 5× figure also remains visible if the advertising report has not been adjusted for refunds. That is why the revenue label should say whether the figure is before or after returns. A correct ratio with an incomplete label can still lead someone towards a poor decision.

Work out the ROAS your margin can support

There is no universal ROAS that proves a campaign is profitable. The required revenue depends on how much each sale contributes towards marketing and the other costs being considered.

In the base example, $27,000 retained revenue minus $16,200 variable fulfilment cost leaves $10,800 before marketing. That is a 40% contribution margin before marketing: $10,800 divided by $27,000.

If that margin stayed constant, every additional dollar of retained revenue would provide 40 cents towards the included marketing costs. To cover $9,000 in marketing, the retailer would need $22,500 retained revenue, because $9,000 divided by 0.40 equals $22,500.

With $6,000 in media spend, the corresponding break-even net-revenue ROAS is $22,500 divided by $6,000, or 3.75×.

This is a threshold for the example’s cost scope and assumptions. It leaves zero contribution after those costs. It does not cover excluded company costs, establish a desirable return or predict what sales will occur at a given advertising budget.

Why the familiar margin shortcut can be too low

You may encounter the formula “break-even ROAS = 1 ÷ margin.” It works for a scope where the relevant margin is available to cover advertising and no other marketing costs are included.

Using the retailer’s 40% margin would produce 2.5×. At $6,000 media spend, that means $15,000 retained revenue. Forty per cent of $15,000 is $6,000, which covers the advertising but leaves the separate $3,000 marketing costs unpaid.

For the example’s fuller scope, the formula is:

Break-even net-revenue ROAS = (media spend + other marketing costs) ÷ (media spend × contribution margin before marketing).

That gives $9,000 ÷ ($6,000 × 0.40) = 3.75×. Reading the dollar amounts first makes the reason for the adjustment easier to understand than memorizing another formula.

If a dashboard reports revenue before refunds, its comparable threshold must allow for the expected refund share. At a constant 10% refund assumption, the retailer would need $25,000 in initial orders to retain $22,500. That is approximately 4.17× gross-revenue ROAS. Keep the refund assumption next to that threshold.

The same ROAS can support very different returns

Suppose the retailer still retains $27,000 but its product mix leaves only a 30% contribution margin before marketing. It now has $8,100 available to cover $9,000 of marketing. The result is a $900 loss, even though its net-revenue ROAS remains 4.5×.

Under that scenario, variable costs are $18,900 and total included costs are $27,900. ROI is −$900 divided by $27,900, or approximately −3.23%.

This comparison is useful when discounts, product mix, shipping or service-delivery requirements change. A revenue dashboard may show little movement while the amount retained from each sale changes materially. Before setting a revenue target, check that the margin used to support it still describes what the campaign sells.

When reports disagree, compare definitions first

An advertising dashboard and a sales ledger can show different totals without either being a simple copy of the other. One report may assign revenue to the date of the advertising interaction; another records the date the order happened. One may include estimated conversions; another contains only recognized orders. Refunds may have been applied to only one of them.

Google’s current conversion reporting documentation distinguishes its primary conversion columns, which use click timing, from “by conversion time” columns. The date range alone therefore does not tell you whether two reports cover the same transactions.

Imagine someone clicks an ad on 29 April and buys on 3 May. An interaction-based April report can include that sale, while an order-date April report will not. Adding a May refund creates another timing difference. Comparing the totals without looking at the dates would turn an ordinary reporting difference into an apparent tracking failure.

The workbook follows a group of advertising interactions and their associated outcomes. The supplied example uses interactions from 1 to 30 April 2026, orders observed through 31 May, and refunds and cost adjustments observed through 30 June. Those dates are illustrative; they are not recommended platform settings.

Choose an observation period that suits the actual sales cycle and the question you are answering. A recent group of leads may still have open opportunities. A recent group of retail orders may still accumulate returns. Label that result provisional rather than comparing it unqualified with a fully observed older period.

Look at how earlier periods developed before deciding when a result is sufficiently settled. If sales from past enquiries continued arriving for several weeks, a report on yesterday’s enquiries should not be treated as complete. If refunds typically appeared later than the initial purchases, a report closed immediately after purchase will systematically miss part of the eventual adjustment.

This does not require waiting indefinitely. You can publish an early operating view and a later reconciled view, provided their status is clear. The early view helps identify activity and obvious problems. The later view supports a more complete cost-and-return comparison. Retain the observation date so a reader can understand why the same campaign’s numbers changed.

For example, a later update that adds sales from the original interactions may improve the reported return without any new advertising. An update that adds refunds may reduce it. Neither movement necessarily reflects a change made to the campaign between the two reports. It may simply be the original activity becoming more fully observed.

Reconcile the difference in amounts

Start with a known source total and explain the differences you can verify: orders outside the date basis, refunds not yet applied, excluded conversion actions or currency translation. Keep any remaining unexplained amount visible.

For the base example, moving from $30,000 before-refund order value to $27,000 retained revenue has an explicit $3,000 explanation. A miscellaneous adjustment entered only to force agreement would make the report less useful.

If two advertising platforms both claim an order, adding their attributed revenue can exceed the business’s sales. Separate platform views can still help understand how each system reports performance. For a combined business calculation, revenue needs a non-overlapping treatment that reconciles to the underlying orders.

Where the records cannot support that reconciliation, describe the gap. More decimal places will not resolve missing order identifiers, incomplete tracking or incompatible attribution rules. A qualified result is preferable to a complete-looking total with an unknown population.

Add campaign amounts before calculating a combined ratio

A monthly report can contain several perfectly correct campaign ratios and still show an incorrect total. The usual mistake is taking their simple average.

For a separate fictional illustration, one campaign spends $1,000 and retains $6,000 in revenue, giving 6× ROAS. Another spends $9,000 and retains $18,000, giving 2×. Averaging 6 and 2 produces 4×, but the business retained $24,000 against $10,000 in media spend. Its combined ROAS is 2.4×.

The larger campaign represents most of the spending, so it must carry more weight in the combined result. Adding the revenue and spending first handles that automatically. A media-spend-weighted average would produce the same answer, but working from dollars is easier to review.

Apply the same principle to ROI. Add the compatible contribution amounts, add the matching cost amounts, then divide total contribution by total cost. An unweighted average of campaign ROI percentages can exaggerate the importance of a small campaign with an unusual result.

Before combining, check that the figures are genuinely compatible. A before-refund revenue ratio and an after-refund ratio should not be pooled without reconciling their revenue. Different currencies need consistent translation. Shared campaign costs must appear only once across the combined total.

A portfolio result also requires non-overlapping revenue. If both campaigns receive full credit for the same order under different reports, adding their amounts will inflate the total regardless of how carefully the final division is performed. Resolve the shared-order treatment first or retain separate reporting views with an explicit limitation.

This is useful in an owner-level report because it lets the campaign detail reconcile to the total being discussed. The total can then answer the business question, while the individual lines explain which activity contributed to it.

Do not carry the fictional answers forward

The workbook is a more detailed version of the comparison above. Its extra fields help keep revenue and costs attached to the same group of advertising interactions. You can read the example without filling them in, but using the model for your business requires replacing the sample data and reviewing its assumptions.

Replace the inputs and cost records

Save a working copy before editing. In Inputs, choose ACTUAL when you have a reconciled schedule of variable costs. Enter gross revenue after discounts but before refunds, then enter refunds as a positive reduction. The model expects this gross-before-refund structure.

Costs flow from CostRegister into the calculation. Replace the fictional cost rows with your own records for variable fulfilment, media, management, creative, measurement and other included marketing. Where a category genuinely has no cost, enter the supported zero-cost record the workbook requires. An unknown amount should remain unresolved.

For a shared cost, the register uses a source amount and an allocation share. It also allows a separate credit. Do not deduct a credit twice by using an already-net source amount and entering the same credit again. Keep the source reference close enough to the entry that another person can check it.

The existing register supports 100 input rows. If your records exceed that capacity, prepare a supported summary or have the model extended and checked before using additional rows. Pasting beneath the supported area is not a reliable way to expand it.

Choose actual costs or a margin scenario

SCENARIO mode estimates variable costs from a contribution-margin assumption. It is useful for asking what the same revenue and marketing costs would mean under a different margin. Keep the scenario label visible when sharing the output.

ACTUAL mode uses the variable costs in the register. The margin input should be blank in this mode. The two approaches are alternatives: subtracting actual variable costs and then applying a margin that already deducts them would count those costs twice.

The Sensitivity tab contains further hypothetical comparisons. These can show which assumption matters, but they do not establish the likelihood of an outcome. A formula responds to what you enter; it cannot tell whether customers will buy, return products or become more expensive to serve.

Read status messages before copying a result

HOLD means a required input or check is unresolved. Review the indicated fields rather than overwriting the output. N/A means a particular ratio cannot be calculated, such as ROAS with zero advertising spend. A numeric zero or a negative return can be a valid result.

READY means the specified checks passed. It does not establish that the source documents are complete or that the campaign caused the sales. Review the workbook’s confirmation fields yourself, including population, timing, cost allocation and currency. Sample YES answers are part of the fictional example, not evidence about your business.

Read the Bridge in dollars before using the ratios in a presentation. The retained revenue should equal variable costs, marketing costs and the contribution left after marketing. Keep the date, revenue basis, cost exclusions and data status with any exported result so its meaning survives outside the workbook.

What the calculation can tell you about spending more

The retailer’s base case leaves $1,800 after the included costs. That is useful information about the selected activity. It does not establish that doubling advertising would double revenue or contribution.

Additional spending might reach different buyers, sell a different product mix or require more creative work. It could also exceed fulfilment capacity. If the next set of orders needs overtime or a new contractor, the next dollar of revenue may have a different cost structure from the previous one.

Evaluate an expansion using the costs and outcomes that would change. Keep that proposed scenario separate from the observed campaign result. An average ROAS describes the whole measured group; a marginal return concerns the additional activity being considered.

Cash timing is another separate issue. Advertising, inventory and delivery may be paid before customers pay. Refunds can arrive after the original sale has made a report look healthy. Positive contribution does not prove that the business can comfortably fund that timing gap.

The figures also describe attributed sales. Attribution assigns credit under a measurement method. Some customers might have purchased without the ad, particularly where they already knew the business or intended to return.

Google’s Conversion Lift documentation describes controlled exposed-and-control comparisons for estimating incremental conversions and notes that the feature is not available to every account. That is a different evidence task from calculating ROAS or ROI on attributed orders.

A spreadsheet cannot settle causation by relabelling attributed revenue as incremental. If the decision depends on how much demand advertising actually added, obtain suitable experimental or other defensible incrementality evidence. Keep its uncertainty visible alongside the economic calculation.

Explain the result in a way someone can act on

A useful summary of the fictional base case would read:

“The campaign was credited with $30,000 in orders and retained $27,000 after refunds. Advertising cost $6,000, giving 4.5× revenue ROAS after refunds. After $16,200 in variable fulfilment costs and $9,000 in marketing costs, $1,800 remained. That is a 7.14% return on the $25,200 of costs included. Company overhead, financing and income tax remain outside this calculation.”

That paragraph answers more than “ROAS was 5×.” It states the outcome, shows where the money went and makes the boundary clear.

The next action should follow the part of the calculation that is uncertain or commercially important. Unreconciled refunds need an order review. An unexplained fulfilment-cost increase needs a product or delivery-cost review. Conflicting revenue totals need measurement work. A campaign-change discussion becomes more useful once those questions are understood.

For campaign tactics after the economics are clear, see the separate PPC ROI improvement guide. Here, the practical finish is simpler: take one recent campaign, calculate retained revenue and the relevant costs, and place those dollar amounts beside its ROAS. The difference between revenue credited and contribution retained will show you what the headline ratio leaves out.

Frequently asked questions

Is ROAS the same as ROI?

No. ROAS divides attributed revenue by media spend, while ROI shows what remains after the costs you include, compared with those costs. A campaign can show an impressive ROAS and still lose money once refunds, fulfilment and other marketing costs are counted.

What is the formula for ROAS and ROI?

ROAS equals attributed revenue divided by media spend. ROI equals revenue minus the included costs, divided by those included costs. Always ask which costs an ROI figure includes, because advertising alone answers a narrower question than advertising, campaign services and the cost of supplying the orders.

Is a ROAS of 4 good?

It depends on the margin the business earns and its other costs. The guide shows how a four times ROAS can still lose money when fulfilment, refunds and marketing services are subtracted. Work out the ROAS the business margin can support before calling any figure good.

What does a 3:1 ROAS mean?

A 3:1 ROAS means each dollar of advertising is credited with three dollars of revenue. Those three dollars also have to cover the ad dollar itself, the cost of what was sold and any other included costs. It describes revenue and leaves profit unanswered.

Does a 300% ROAS mean a 300% profit?

No. A ROAS shown as a percentage is the same revenue ratio as a multiple, so 300% ROAS and three times ROAS are identical. The revenue still has to pay for the advertising, the product or service and any other costs before anything is left over.

How is ROAS different from customer acquisition cost?

ROAS looks at attributed revenue for each advertising dollar. Customer acquisition cost looks at what you spend to win one new customer. They answer different questions, so read them alongside an estimate of what a customer is worth over the relationship.

How does ACoS relate to ROAS?

ACoS divides ad spend by attributed sales, which makes it the inverse of ROAS, so a 25% ACoS matches a ROAS of four. A lower ACoS means more revenue per ad dollar. Like ROAS, it says nothing about margin or the other costs that decide whether a campaign makes money.

What break-even ROAS can my profit margin support?

Start with the revenue you actually keep after refunds, then put back fulfilment, management, creative and measurement costs. The familiar shortcut of dividing one by the margin can set the target too low, because it ignores costs beyond the media. The same ROAS can support very different returns for two businesses with different costs.