True ROAS vs platform ROAS: how to find what your ads really return

True ROAS is return on ad spend calculated with the revenue your business actually recorded, not the revenue the ad platforms attribute to themselves. The inputs come from your shop system, your CRM or your accounting, and the spend is the sum of all ad platforms in the same period.

True ROAS = actual revenue (shop or CRM) ÷ total ad spend

The term overlaps with two others. MER (marketing efficiency ratio) divides total company revenue by total marketing spend. Blended ROAS usually means the same calculation limited to paid media. In practice all three answer one question: does the money going into ads come back as real revenue? Shopify describes MER the same way, as total revenue divided by total marketing spend (Shopify, 2026).

Each ad platform only sees its own ads and credits itself whenever one of them touched the customer. None of them knows that another platform also claimed the same order. Add the reports up and the total exceeds what you actually sold.

• Overlapping attribution windows. Meta's standard attribution can credit purchases up to 7 days after a link click and 1 day after a view (Meta Business Help Center). Google Ads uses a 30-day click window by default (Google Ads Help). A customer who clicks a TikTok ad, sees an Instagram ad and then searches your brand on Google can appear in all three reports.
• View-through conversions. A purchase after someone merely saw an ad counts, even if they would have bought anyway.
• Data-driven attribution only splits credit inside one platform. Google's data-driven model distributes credit across Google ads, not across Meta or TikTok (Google Ads Help).
• Modelled conversions. When users decline tracking, platforms estimate the conversions they could not observe (Google Ads Help). The estimate lands in your report next to real orders.
• Existing demand. Branded search and retargeting often reach people who were already on their way to buy. The platform gets the credit, the ad did little.

Research backs this up. A study of 15 large US advertising experiments at Facebook compared common observational measurement methods with randomised tests and found that the observational methods often failed to recover the effect the experiments measured (Gordon, Zettelmeyer, Bhargava and Chapsky, Marketing Science, 2019).

A connected car app has run its advertising with the same agent system behind Marketing AIgency since 14 September 2026, across 9 markets on Meta, Google and TikTok, with about €470,000 in monthly ad spend under management. For Germany in September 2026 we put the platform reports next to the revenue the company recorded.

Google alone claimed more revenue than the whole company made in that market. Together the platforms claimed 2.2 times the actual revenue, so the over-counting factor was 2.2. The Meta and TikTok row is derived from those two published figures, not a separate report.

The factor translates directly into ROAS. Whatever the spend, if the platforms together report a combined ROAS of 4.4, the true ROAS at a factor of 2.2 is 2.0. For a company with a 40% gross margin, that is the difference between scaling happily and sitting 20% under break-even. You can check your own break-even with our ROAS calculator.

MER = total company revenue ÷ total marketing spend

Blended ROAS = total revenue ÷ total paid media spend

Over-counting factor = (Meta revenue + Google revenue + TikTok revenue as reported) ÷ actual revenue

The calculator at the top of this page takes your total ad spend, the revenue each platform reports and your actual revenue for the same period. It returns each platform's ROAS, your true ROAS or MER, and the over-counting factor. A factor of 1.0 would mean the platforms agree with your books. The further it sits above 1.0, the more optimistic any budget decision based on platform ROAS becomes.

1. Pick the same date range in every platform and in your shop. Use calendar months to avoid time-zone gaps.
2. Export spend and reported purchase value per platform.
3. Take net revenue from your shop or CRM: after VAT, returns and cancellations.
4. Decide whether total revenue includes organic and email sales. MER does, a paid-only view does not. Stay consistent month to month.

The company number is the revenue or lead count that exists in your own systems. Getting it is mostly a data plumbing job, and each method below adds a layer of accuracy.

1. Shop or CRM revenue. Shopify, WooCommerce or a CRM export gives you the total. It tells you what happened, not which ad caused it.
2. First-party pixel. A tracking script on your own domain records visits and orders with the click IDs and UTM parameters of each ad. It sees all channels at once and counts each order once.
3. Post-purchase survey. One question after checkout, such as how did you hear about us, captures channels no pixel can see: podcasts, word of mouth, a TikTok someone watched on another phone.
4. Incrementality tests. Meta's Conversion Lift splits an audience into exposed and holdout groups (Meta). Geo tests switch ads off in some regions and compare, for example with Meta's open-source GeoLift (GitHub).
5. Marketing mix modelling. For larger budgets, models such as Google's open-source Meridian estimate each channel's contribution from spend and revenue over time (GitHub).

Most small and mid-sized advertisers get far with the first three. Incrementality tests need enough volume and a few weeks of patience, so they fit best when you are about to make a big budget shift.

Use MER or true ROAS to decide how much to spend in total, and platform ROAS to decide where to spend it inside a platform. Platform numbers are biased, but the bias is fairly similar across campaigns within one account, so they still rank ad sets and creatives reasonably well.

• Raising or cutting the monthly budget: MER or true ROAS against your break-even.
• Shifting budget between Meta, Google and TikTok: first-party pixel data and survey answers, confirmed by a test if the shift is large.
• Pausing an ad or scaling an ad set: platform ROAS or CPA, read against the over-counting factor for that platform.
• Judging a new channel: a holdout or geo test, because its platform report has no track record yet.

Marketing AIgency's agents steer on the company number instead of platform ROAS. Revenue comes from Shopify, a CRM, a Google Sheet or the product's own first-party pixel, and a post-purchase survey adds the channels a pixel cannot see. Purchases can be sent back to Meta through the Conversions API. Every suggested change lands in an approval queue with the data and reasoning behind it, and its effect is measured after 7 days.

If you currently use a dedicated attribution tool, our comparisons with Triple Whale and Hyros show where each fits better. Plans have fixed monthly prices with no percentage of ad spend, see pricing.