The report shows sales, and the amount attributed to ads looks convincing. Yet the company's account doesn't hold what you expected. That isn't necessarily a contradiction: the advertising report and the economic result answer different questions. The first links revenue to advertising under certain rules. The second must also include the costs that make the sale possible.
A profitable ROAS depends on the contribution of the products, operating costs, returns and the financial goal. First calculate how much remains of net revenue before ads, then decide how much you can allocate to advertising. Compare metrics on the same basis and separate attributed revenue from retained orders. A high ratio between revenue and media doesn't guarantee net profit.
The priority is to define the revenue, the spend and the period. Only then does it make sense to compare campaigns, set a target or discuss increasing the budget.
Contents
- What ROAS measures
- The costs that change the result
- Three hypothetical calculations
- The target in the platform
- Categories, returns and new customers
- The decision report
- Practical plan
- Frequently asked questions
- Checklist
What ROAS measures and what it leaves out of the calculation
ROAS is the ratio between the revenue attributed to ads and the advertising spend. The simple formula is ROAS = attributed revenue / media cost. A value of 4 means four units of attributed revenue for each one spent on media. It can also be expressed as 400%, if you keep the same definition.
The metric doesn't automatically subtract the cost of products, shipping paid, management or salaries. Nor is attributed revenue always identical to the net revenue of retained orders. That is why two stores with the same ROAS can have very different economic results.
Define the basis before comparing
State in the report whether revenue includes VAT, delivery, discounts and orders later cancelled. Specify whether the media includes taxes or relevant currency conversions. You don't need an identical definition for all the reports in the company, but you must be able to reconcile the differences between them.
A campaign that reports gross revenue can look more efficient than one analysed after returns. The difference doesn't prove better performance. Before the decision, bring the data to a comparable basis and keep the adjustments made visible.
Attribution is not causality
Platforms decide which interactions they attribute an order to, using their own windows and rules. The same transaction can be claimed by two systems. Don't automatically add up the revenue attributed by Google and Meta as if it represented distinct orders.
A customer can discover the product in a video, later search for the store's name and buy after an email. A channel's report describes one perspective on the path. To evaluate the business you also need the shared order record, plus a cautious interpretation of the contribution of each activity.
The costs that change the economic threshold
Contribution before ads
Start from the relevant net revenue and subtract the variable costs. Include the cost of goods, processing, packaging and shipping paid, if they are tied to orders. Treat returns consistently: either adjust the revenue and the actual costs, or use an explicit estimate until the period matures.
The result is the contribution available before advertising. From it you must fund the ads and, after that, fixed costs and the desired result. If ads consume the entire contribution, the store can generate volume without covering the rest of its operation.
Margin is not markup
If a product costs 100 RON and is sold for 150 RON on the same tax basis, the difference is 50 RON. The markup relative to cost is 50%, and the margin relative to price is about 33.33%. Confusing the two percentages changes a profitability calculation strongly.
For the ads ceiling, use a contribution defined after the relevant costs, not a markup percentage taken from inventory. Check the formula with the person who manages finances. A simple correct calculation is more valuable than a detailed dashboard built on wrong definitions.
Fixed costs remain real
Rent, salaries, software and other costs don't disappear because they are not in the ROAS formula. You can first analyse the campaign's marginal contribution, but don't automatically call it net profit. Show separately what remains for fixed costs and how it relates to what the business needs.
Management and the production of materials must be treated consistently. You can have a media metric and a full marketing metric. Comparing them is useful if the labels are clear; alternately using whichever definition looks better creates confusion.
Three hypothetical examples that change the interpretation
Hypothetical example 1: the simplified threshold and the working target
For a category, in a completed month, you have net revenue of 30,000 RON and variable costs of 21,000 RON, before ads. The contribution is 9,000 RON, that is 9,000 / 30,000 × 100 = 30%.
If you were to spend all the contribution on media, the simplified threshold would be 30,000 / 9,000 ≈ 3.33 ROAS. Nothing would remain of this contribution for fixed costs or profit. The equivalent formula is 1 divided by the contribution rate, expressed as a decimal: 1 / 0.30.
If you want to keep 4,000 RON, the maximum media in this model is 5,000 RON. The corresponding ratio becomes 30,000 / 5,000 = 6. The working target therefore differs from the threshold that consumes the entire contribution. Both are calculations for the given assumptions, not market standards.
Hypothetical example 2: the same ROAS, different contribution
Two categories each produce 20,000 RON of attributed net revenue and each consume 4,000 RON of media in a shared period. Both have a ROAS of 5. Category A has a contribution before ads of 40%, that is 8,000 RON. Category B has a contribution of 20%, that is 4,000 RON.
After media, A keeps 4,000 RON and B nothing, before fixed costs. An aggregate report doesn't show you this difference on its own. If expansion shifts sales towards category B, revenue can rise without a corresponding rise in the contribution that remains.
The useful action is to analyse the product selection and the economics of the categories. Don't change prices or targets just to get the same ROAS everywhere. Some categories can have distinct commercial roles, but these must be explained and evaluated, not assumed.
Hypothetical example 3: reported revenue and retained orders
In a 30-day test, the platform attributes 24,000 RON of revenue and 4,000 RON of media: ROAS 6. After cancellations and returns are clarified, the relevant revenue of retained orders in the scope analysed is 20,000 RON. The adjusted commercial ratio is 20,000 / 4,000 = 5.
If you add another 1,000 RON for materials and management, the ratio between adjusted revenue and total cost becomes 20,000 / 5,000 = 4. This has a different denominator and must be named explicitly, without claiming it is the same column in the platform.
Suppose there are 80 retained orders, from 60 new customers and 20 existing ones, one per person. The media cost per order is 50 RON, and allocating the whole cost to new customers is about 83.33 RON per customer. The definitions answer different questions and shouldn't be mixed.
The ROAS target in the platform is not the profit calculation
A bidding strategy can use conversion values to pursue a reporting goal. Google explains the Target ROAS strategy. The setup must take into account the data available, the campaign type and the values sent. A goal set is not a guarantee of results.
Check the values sent
If a purchase event sends the same value for all orders, commercial differences can be lost. If the value is duplicated, the system can receive a wrong picture. Start with a test order and follow the identifier, currency and value into the reports.
In GA4, using the transaction ID correctly helps to deduplicate web purchases. This doesn't automatically solve all problems in all platforms. The integration must be verified in each system that receives events.
Don't ask for more than the model can sustain
A very ambitious target doesn't create margin and doesn't improve the offer on its own. If the desired result is incompatible with the costs and demand observed, you may see reduced volume or unstable results. Analyse the economic cause before you treat the setting as a universal solution.
In evaluating Google Ads campaigns, ask for an explanation of the link between goal, data and products. For Meta campaigns, request the same discipline of definitions, even though the interface and options differ. The platform doesn't replace the decision about acceptable contribution.
Read the result by category and customer type
Returns can change the order of the winners
A category with high initial sales can have additional costs after delivery. If you compare it too early with a stable one, you may favour exactly the selection that leaves less contribution. Choose periods mature enough for your commercial process and mark results that are still provisional.
Don't use the same estimated return rate for all products if you have evidence of important differences. On the other hand, don't build precise percentages on two or three orders. Keep the level of detail proportional to the volume and update the estimates when better data appears.
The new customer can have a different economic role
A channel that sells predominantly to existing customers can have a good ROAS without significantly expanding the buyer base. Another can bring new customers at a higher initial cost. Compare the roles, but don't use the customer's future value as an automatic excuse for present losses.
For repurchase, track groups of customers and later contribution, not just revenue. Include the time to return and the retention costs. A promise of long-term value is different from a behaviour observed in the store's data.
The buying experience counts in the calculation
An incompletely described product can generate orders followed by dissatisfaction and return. Improving the page can change both conversion and the quality of orders. If the store needs Shopify development, tie the intervention to an observed problem, such as choosing the wrong variant.
Don't track only the increase in the order rate. A page that clarifies incompatibility can reduce some unsuitable purchases and can improve the final contribution. The good decision shows in the whole commercial result, not in a single percentage.
A report that helps the decision
| Metric | The question it answers | Limit of interpretation |
|---|---|---|
| Platform ROAS | What revenue does the system attribute to ads? | Depends on attribution and values |
| Retained net revenue | What sales remained after adjustments? | Requires orders to mature |
| Contribution before media | What economic room do we have? | Depends on full costs |
| Contribution after media | What remains after distribution? | Not automatically net profit |
| Cost per new customer | How much do we allocate to expanding the base? | Requires consistent identification |
| Cash | Can we sustain the commercial cycle? | Cannot be deduced from ROAS |
Add the period, currency and sources next to the table. Include a column with the important changes: price, offer, stock, tracking, shipping. If the result changed after three simultaneous interventions, say that you can't precisely isolate the contribution of each.
The report should end with a decision and an open question. For example, keeping the investment in one category and checking returns before expanding. A list of metrics without commercial consequences shifts the interpretation onto the entrepreneur, instead of helping them.
A practical plan for a correct analysis
In the first days, define the revenue and the costs together with the person responsible for finance. Choose a mature period and reconcile the orders. Note the differences between the store and the platforms without assuming that all of them are errors; some come from different rules of attribution and timing.
Then calculate the contribution for the important categories and build two thresholds: the simplified limit and the target that keeps the necessary reserve. Check the sensitivity to discounts, returns and a change in mix. A single average calculation isn't enough if the products differ a lot.
In the next stage, review the values sent and the goals the campaigns use. Decide on one justified intervention and document the moment. Don't change the report's definitions in the same period without a clear note, because you will lose comparability.
At the end of the cycle, compare the observed result with the hypothesis, taking conversion delay into account. Decide what to continue, what to reduce and what information is missing. If you need help, talk to BlueAds starting from this set of data.
Frequently asked questions
Is a ROAS of 4 good?
It can be or not, depending on contribution, returns and the remaining costs. There is no universal answer. Calculate what remains after variable costs and media, then check whether the sum can support the operation and the company's goal.
Are ROAS and ROI the same thing?
No. ROAS relates attributed revenue to media. ROI evaluates the profitability of an investment based on the result and the cost defined for that investment. State the formula used; labels alone don't ensure comparability between reports.
Can I directly compare Google ROAS with Meta ROAS?
Only with attention to windows, rules, values and the role of the campaigns. Orders can overlap, and audiences can have different intents. Also use the shared commercial record, not just two columns from different interfaces.
What do I do if ROAS falls after I increase the budget?
Check the maturity of the data, the product mix, the additional demand and the absolute contribution. A lower average doesn't prove on its own that expansion is wrong. What matters is what economic result the additional money brings and whether the risk remains acceptable.
Do I include returns?
For commercial analysis, their effect must be reflected consistently. You can use mature data or labelled estimates for recent periods. Don't compare an adjusted month with an unadjusted one without explaining the difference.
Can I set the calculated threshold directly as the target in the platform?
Not mechanically. Check the basis of the revenue reported, the conversion values and the conditions of the strategy. A net economic threshold can have a different definition from the value used for bidding. Align the model and the data first.
Can a high ROAS hide stagnation?
Yes, if it comes predominantly from demand that already existed or from customers who were coming back anyway. Also analyse new-customer acquisition, volume, contribution and the development of demand. Reported efficiency and business growth are related but distinct questions.
Checklist for a useful ROAS conversation
- Revenue and spend have explicit definitions.
- Returns and cancellations are treated consistently.
- Margin is not confused with markup.
- Contribution is separated from net profit.
- Important categories are analysed separately.
- New customers are not confused with all orders.
- Conversion values have been verified.
- The decision takes cash and capacity into account.
ROAS is useful when you put it in context. Use it to understand advertising reporting and complete it with the store's real economics. That way the conversation about performance can lead to a commercial decision, not just a prettier number in a report.