What is ROAS, how to calculate it and what counts as a good ROAS

According to Google Ads Help, ROAS compares the reported conversion value with what you spend on ads. That reported value is not necessarily the revenue your online store actually realizes.

Data and settings described in this article are valid as of November 10, 2026.

ROAS equals conversion value divided by ad spend. In Google Ads, you will find it as the Conv. value / cost metric (Google Ads Help).

ROAS can be expressed as a multiple or as a percentage. A ROAS of 4× is the same as 400%. Before you interpret a number, check which unit the report uses (Google Ads Help).

How ROAS, PNO, ROI, and POAS differ

Each of these metrics uses different inputs. We recommend clarifying exactly what goes into the numerator and denominator before comparing them.

The definitions used by Google and Sklik allow the following conversion only when both metrics use the same underlying data.

The equation ROI = POAS − 1 would require matching definitions of profit and the denominator, which the metric names alone do not ensure (Google Ads Help).

Two stores can report the same ROAS and still have different economic outcomes when their margins differ. That follows from the OpenStax contribution margin model, with related economic context discussed by Humblytics.

What is left before you pay for ads

We use “pre-ad contribution” to describe the amount remaining from revenue after relevant variable costs, before advertising is deducted. The “pre-ad contribution margin ratio” expresses that amount as a share of revenue. These descriptive terms apply the OpenStax contribution margin principle, keeping the monetary amount separate from its ratio.

Guides on ad economics note that, besides the purchase price, shipping paid by the store and payment fees may need to be included (Humblytics, Polar Analytics). We recommend also considering packaging, fulfillment, and the cost of returns. Which costs belong in the calculation depends on how your business is structured (OpenStax).

As OpenStax explains, products can differ in both their contribution amount and contribution margin ratio. An overall ROAS can conceal those differences between product groups. Under the Google definition, ROAS itself deducts neither other variable costs nor fixed costs, leaving those parts of the economic picture outside the metric.

Separate ad-cost coverage from your business target

Let m be the positive share of revenue remaining as pre-ad contribution, entered as a ratio. The following derivation applies the OpenStax principle and requires a consistent data basis throughout.

• Ad-cost break-even ROAS = 1 / m, under those conditions.

• Corresponding PNO (%) = 100 × m, using the same model.

This is a conditional mathematical derivation built on the contribution model and the definition of ROAS (OpenStax, Google).

Covering advertising does not automatically cover fixed costs or provide the profit you require. A contribution amount calculated before fixed costs is not net profit. See the OpenStax explanation and the target-setting approach in the PNO calculator.

You can derive your own target from this model and from the PNO calculation.

A worked example

The inputs below are assumptions we chose for illustration. OpenStax documents the contribution principle, not the specific values in this example.

Table source: OpenStax, Sklik Help.

An archived request from one client

An anonymized archive record from February 2026 concerns one online store. The client linked the PNO they wanted to the margin they reported for their business. The archive record documents this client’s request and their description of the situation at the time.

According to the same archive record, our team emphasized the importance of margin and pointed out that promoting only expensive configurations could limit both the conversion rate and the volume of data available to campaigns.

Our take: Do not copy your target ROAS from a benchmark or a competitor; calculate it from your own margin. Triple Whale reports a median Google Ads ROAS of 3.27× for the brands in its report. A margin of 13 to 16% with a PNO of 9% leaves just a few percentage points, before checking whether shipping, payments, and returns are included.

A store like this first needs an exact figure for what it keeps from each order. Do not force PNO down by promoting only expensive configurations: this could limit the data available for bidding to learn from.

Measurement documentation stresses that advertising costs and order values must be treated consistently for value-added tax (VAT), and it flags differences in reporting periods and attribution. Attribution means assigning results to advertising. Converting ROAS into PNO cannot resolve those differences. See Sklik Help, Heureka Help, and Google Ads Help.

Czech VAT, cancellations, and returns

Sklik reports advertising prices excluding VAT (Sklik Help). Heureka lets you set up conversion tracking either with or without VAT, and it requires costs to match the order values you send (Heureka Help).

For its conversion tracking service (Měření konverzí), Heureka assumes by default that order values include VAT, according to Heureka Help. The setting must match the values transmitted by the store.

In the Czech Republic, as of November 10, 2026, the standard VAT rate is 21% and the reduced rate is 12%. Section 47(1) of Act No. 235/2004 Sb., in the version effective in 2026, sets out these rates. The same rate does not apply to all goods. See the Czech VAT Act and the Czech Financial Administration.

Google Ads lets you adjust recorded conversions after the fact: you can restate a conversion’s value or retract the conversion entirely (Google Ads Help).

The following relationships are derived from the documentation on conversion adjustments and the accounting treatment in OpenStax. Let r be the share of originally reported revenue that was later returned. Ad spend stays the same, revenue is measured on the same basis, and 0 ≤ r < 1.

ROAS after returns = original ROAS × (1 − r).

Relative overstatement of original ROAS versus corrected ROAS = r / (1 − r).

The relative comparison uses corrected ROAS as its denominator. See Google Ads Help and OpenStax.

The OpenStax accounting explanation shows why revenue and profit need separate treatment. A return reduces revenue, while returning merchandise to inventory also reduces the cost of goods sold. The share of value returned therefore cannot, by itself, determine the effect on profit.

As of the verification date, Sklik’s public documentation does not describe a general procedure similar to the one in Google Ads (Sklik Help, Sklik Help, Google Ads Help). By that procedure we mean restating the value of, or retracting, a specific recorded conversion after the fact.

Why results can differ between reporting systems

Google Ads, Google Analytics 4 (GA4), and the store’s own system can apply different attribution rules and assign results to different dates. Google Ads may report a conversion on the date of the ad interaction, while an analytics comparison may use the conversion date. The relevant rules are described in Google Ads Help, Google Ads Help, and Sklik Help.

Several platforms can claim the same order (Sklik Help, Heureka Help). Adding up the values each platform reports therefore does not automatically give you unique revenue.

A historical attachment in our communication with one online store showed the same ad set in Ads Manager for the period from February 1, 2025 to April 30, 2025. It displayed the PNO under two attribution settings, 1-day view and 7-day click. It illustrates how platform-reported PNO depends on attribution.

For related context, read our article on performance marketing and how to evaluate it.

tROAS stands for Target ROAS. In Google Ads, automated bidding uses reported conversion values and aims for the average ROAS you set as a target (Google Ads Help). According to Google Ads documentation, individual conversions can differ from the target average (Google Ads Help).

Moving to Target ROAS in Google Ads

Google distinguishes the technical minimum from the recommended history of reported values.

Source: Google Ads Help, Google Ads Help.

What Sklik’s public documentation establishes

Sklik Help documents Maximalizace konverzí (a conversion-maximizing strategy with max. CPA, meaning a maximum cost per conversion). It lists Maximalizace obratu podle max. PNO (revenue maximization based on maximum PNO) as planned development. The Sklik terms mention revenue maximization, while the help documentation lists it as planned development.

Maximalizace konverzí requires at least 30 conversions in the last 30 days, and the campaign has to be active for at least 7 days before you turn it on (Sklik Help). Sklik’s documentation specifies these conditions for the max. CPA variant. The contractual terms that mention revenue maximization took effect on April 1, 2026 (Sklik terms).

Triple Whale provides an attribution tool. Its published methodology describes medians, cohorts (groups of brands), and currency normalization (Triple Whale methodology). Triple Whale’s benchmarks describe this provider’s customer groups during specific periods (Triple Whale).

Median ROAS in Triple Whale reports

We take the values directly from Triple Whale’s own reports.

At PPC ONE, we recommend resolving the following questions before calculating a target. This checklist follows from the metric definitions and conversion tracking documentation.

With consistent inputs, calculate ad-cost break-even ROAS and then your own economic target. Assess the bidding strategy’s technical requirements separately. Also allow for conversion lag, the time between an advertising interaction and a conversion. This sequence draws on the contribution margin principle and Google Ads Help.

If conversion value cannot be explained or margin is defined inconsistently, request supporting detail before judging performance. PPC ONE’s recommendation is based on the economic model and conversion tracking documentation (OpenStax, Sklik Help).

We provide services for Google Ads, Facebook Ads, Instagram Ads, and Seznam Sklik, along with server-side tracking and cookies. Explore our services (PPC ONE).