ROAS is useful for monitoring ecommerce advertising, but it is not enough to decide whether a campaign deserves more budget. It measures attributed conversion value against advertising spend. It does not, by itself, tell you how much money remains after fulfilment, whether buyers are new, or how many sales would have happened without the advertising.
A dashboard showing 5x ROAS can look reassuring while the underlying orders leave very little contribution. Equally, a lower ratio can accompany an increase in profitable sales. The decision depends on what the reported value includes, the economics of the products sold and the role of the campaign. This guide gives store owners and marketing teams a practical way to connect advertising reports with business decisions. All numerical examples are hypothetical, not WeAreMedia client results or industry benchmarks.
Establish what your ROAS number actually contains
The basic calculation is ROAS = attributed conversion value / advertising spend. If the value sent to the platform is revenue, spending £2,000 and reporting £10,000 in attributed sales produces 5x ROAS, or 500%. It does not establish that the business made £8,000 in profit.
Conversion value is an input to reporting and bidding. It may represent sales revenue or another business value, depending on the implementation. Google’s explanation of conversion values describes that distinction. Read the configuration rather than assuming that every column labelled “value” measures the same thing.
Write a shared definition before comparing reports. Does the value include tax or shipping charges? Is it before or after discounts? Which currency is used? When are cancellations and refunds reflected? Also distinguish orders from customers: three orders from one person are not three newly acquired customers. These definitions matter more than an attractive headline ratio.
A store selling in several countries needs consistent treatment of currencies, reporting periods and costs. A difference in exchange rates or delivery subsidies can change the economics even when the campaign ratio appears stable. Keep those assumptions visible in the report so that a later comparison does not silently change its basis.
Calculate break-even from contribution, not product markup
For this guide, contribution before advertising means net sales minus the variable costs needed to supply those sales. Those costs can include the product, payment processing, packaging, fulfilment and the shipping cost absorbed by the business. Advertising and fixed overheads have not yet been deducted.
Hypothetical order: Net revenue, excluding tax and after discounts and refunds, averages £100. The product costs £45, subsidised shipping £8, payment processing £3, and packaging plus variable fulfilment £4. Total variable costs are £60. The remaining contribution before advertising is £40, or 40% of net revenue. Any additional return processing or unrecoverable inventory costs also need an explicit place in the calculation.
Break-even ROAS = 1 / contribution margin before advertising. At a 40% contribution margin, the ratio is 1 / 0.40 = 2.5x. On a consistent revenue basis, 2.5x covers those variable costs and the advertising. It leaves nothing for fixed overheads or an operating profit target.
If you want 10% of net revenue left after advertising, the simplified example leaves 30% available for media. The corresponding ratio is 1 / 0.30, approximately 3.33x. That remaining 10% is still contribution, not automatically net profit. Rent, fixed salaries and other overheads still need to be covered.
Do not apply this threshold directly to a dashboard reporting tax-inclusive order totals when your margin calculation excludes tax. The numerator must be comparable. Recalculate when product mix, discounts or delivery costs change. A break-even figure is a decision input based on assumptions, not a permanent property of the account.
The same ROAS can mean opposite things for two product groups
Imagine one product group has a 20% contribution margin before advertising and another has 50%. Both report 4x ROAS on comparable net revenue. For every £100 sold, advertising costs £25. The first group contributes £20 before ads and loses £5 after them; the second leaves £25 after advertising.
A store-wide average can therefore hide weak economics in one category. Group products by factors that materially change the decision: margin, return exposure, shipping weight, discount dependence or repeat-purchase behaviour. You do not necessarily need a separate campaign for every group. Start with useful reporting distinctions and sufficient sales volume to interpret them.
Avoid splitting a small account into so many segments that one order changes the entire conclusion. The purpose is to identify where additional spend has a reasonable economic basis, not to manufacture a league table of tiny samples.
Packaging belongs in this discussion too. A cheap box can become expensive if it adds packing time, damage or replacement shipments. Our guide to ecommerce packaging and shipping protection explains how to assess those costs at order level instead of looking only at the purchase price of the box.
Treat refunds and delayed conversions as part of the reporting process
Advertising spend appears quickly, while the final value of an order can take longer to settle. Cancelled orders, partial refunds and failed deliveries can change the picture. Keep a current operational view and a separate review of order groups whose returns have had time to emerge.
Suppose a campaign initially reports £10,000 in order value from £2,000 of spend. If comparable net sales for those orders later settle at £8,000, the revenue ratio falls from 5x to 4x. Return shipping and handling may reduce contribution further. The early report was provisional; it was not necessarily a reliable statement of final profitability.
Google Ads supports restating or retracting eligible conversions, with requirements that depend on the measurement setup. Check the conversion adjustment documentation before designing the process. A refund in the commerce platform does not prove that every advertising report has been updated automatically.
Avoid subtracting the same refund twice. If refunded sales have already been removed from net revenue, do not also record the whole refund as an additional cost. Track genuinely additional costs, such as return handling or inventory that cannot be resold, under their own definitions. Agree the treatment with the person responsible for the business accounts.
Recent campaign results can also be incomplete because purchases happen after an advertising interaction. Google’s conversion lag report helps identify that delay. Choose an evaluation period that allows meaningful comparison rather than treating yesterday’s ratio as a settled result.
Reconcile platforms without adding their claimed revenue together
A customer might discover an item through social advertising and later click a search ad before ordering. Different platforms can assign credit to the same transaction under their respective rules. Adding their attributed revenue together does not produce the store’s unique sales total.
Attribution distributes credit among interactions. It does not, on its own, establish that an ad caused a sale. GA4 reporting models and lookback settings affect which interactions are considered; the Analytics attribution settings documentation explains the available framework.
Use the commerce system’s unique orders and realised net sales as the business-level reference, alongside total media spend. Use individual advertising platforms for diagnosis within their own measurement framework. When numbers differ, check time zones, currencies, revenue definitions, reporting dates, consent coverage and refund cutoffs before declaring tracking broken.
The reports do not need to agree perfectly to be useful. They need clearly explained differences. For a new store, establishing those definitions before launch is much easier than reconstructing them after several campaigns are running. Our Shopify launch planning guide connects measurement with the wider operational decisions that need to be ready before paid traffic starts.
Separate acquisition from repeat purchasing
Returning customers are valuable, but a campaign reaching people who already know the brand has a different job from one introducing it to new buyers. Review new and existing customers separately where the data supports that distinction. Do not treat a different device or browser as proof of a new customer.
New-customer acquisition cost divides a defined acquisition spend by the relevant number of new customers. If you include only media, label it that way. A broader calculation including creative production, agency fees and relevant staff costs answers a different question. Keep both definitions stable when comparing periods or channels.
For a wider view, divide the store’s net revenue by total paid media spend over the same period. Here, we call that the blended revenue-to-media ratio. Its numerator includes sales influenced by email, organic search, direct visits and repeat customers. It is useful for observing the whole business, but it does not attribute all revenue to advertising.
Repeat purchasing can justify a different acquisition approach, provided the evidence and cash requirements support it. A recently launched store does not yet have a proven annual customer value. Track realised contribution from customers acquired in the same period over comparable 30-, 60- or 90-day windows. Show future value as a forecast rather than using an unlimited lifetime estimate to explain away current losses.
Consider what the next increment of spend might achieve
Average ROAS describes the spend already made. It cannot guarantee that the next £1,000 will perform at the same rate. Broader audiences, changed product mix or different buying intent can alter the return from additional spend.
Hypothetical comparison: One period produces £10,000 net revenue from £2,000 media spend. A comparable later period produces £13,000 from £3,000. Average ROAS falls from 5x to about 4.33x. The observed additional revenue divided by additional spend is £3,000 / £1,000 = 3x. At a 40% contribution margin, that additional revenue creates £1,200 before advertising, leaving £200 after the extra spend.
This comparison is not a causal experiment. Seasonality, prices, stock, other channels and returns may have changed. It simply shows why a lower average ratio can accompany higher total contribution. Check total contribution and customer acquisition before concluding that a falling ROAS means the campaign has deteriorated.
For a budget test, specify the spend limit, review date and conditions for stopping. Changing prices, targeting, campaign structure and the website simultaneously makes interpretation harder. If the test generates too few orders, record that uncertainty instead of presenting a small movement as a dependable result.
Audit the measurement inputs before changing a bidding target
Before increasing spend, also check conversion counting, search intent and the action on the landing page. Our guide to ten Google Ads campaign mistakes helps separate measurement and traffic problems that can distort a ROAS review.
A tidy dashboard is not evidence that purchases are recorded correctly. Reconcile a sample of authorised order records with the measurement implementation. The technical team should do this without unnecessarily sharing customer information.
- Does the purchase event correspond to a confirmed order?
- Does the transaction identifier help prevent repeated counting of the same order?
- Do value and currency match the intended revenue definition?
- Are test orders, cancellations and refunds handled consistently?
- Could multiple tags or primary conversion actions count the same purchase twice?
- Does measurement respect the visitor’s consent choices?
- Does the report state its cutoff and the possibility of delayed conversions?
A coverage gap is not a reason to override consent. Explain the limits of the available data. If unexplained discrepancies remain, scope a conversion tracking and analytics review around event accuracy, order reconciliation and reporting definitions.
Google Ads Target ROAS bidding works with the conversion values supplied to it. A very high target can restrict the traffic available to a campaign; entering a higher number does not fix poor product economics. The Target ROAS documentation describes this behaviour. Set expectations from verified measurement and business constraints, then allow the chosen evaluation period to complete.
Ask whether the advertising creates additional demand
Attribution cannot fully separate orders caused by advertising from orders that might have happened anyway. This deserves particular attention for branded search and remarketing. But turning a campaign off for a few days and comparing revenue is not automatically a reliable incrementality test.
Where volume permits, a controlled comparison can investigate additional sales or contribution. Decide the outcome, sample requirements, test duration and potential interactions with other channels in advance. Groups must be sufficiently comparable for the result to be useful. Promotions, regional differences and existing customer behaviour can otherwise distort the conclusion.
A small store may not have enough volume to detect a dependable difference. In that case, acknowledge the uncertainty and combine store-level results, customer cohorts and campaign data. Do not claim that a channel creates no additional demand merely because an underpowered test produced an inconclusive result.
Make the weekly report end with a decision
A report that ends with “ROAS was 4.2x” leaves the important questions unanswered. Begin with the period and revenue definition. Explain the main change, identify the evidence supporting it and state the next action.
- Did net sales, media spend and contribution after advertising move together?
- Which product group and customer type account for the change?
- Did returns, discounts, availability or delivery conditions affect the outcome?
- Is any measurement difference or conversion delay still unresolved?
- What single assumption will the next test examine, and when will it be reviewed?
High ROAS with declining new-customer numbers may call for an acquisition review. Lower ROAS with positive contribution and growing customer acquisition may support a different decision. Rising revenue with weak contribution can point to pricing, product mix or fulfilment costs. The same ratio should not trigger the same recommendation for every store.
When discussing Google Ads management with WeAreMedia, bring product-group costs, return data and measurement definitions alongside the campaign reports. That makes it possible to define a useful scope: which sales are worth growing, what remains uncertain and which controlled next step can answer the question.