The ROAS lie every webshop tells itself

Your Google Ads dashboard shows 8x ROAS. Meta shows 4x. Email shows 20x. You think: great, scale everything.

Then you check your overall profitability and it barely moved. Why?

Because blended ROAS hides where revenue actually comes from. And attribution platforms disagree by 40%+ on the same data. Your ROAS dashboard is not lying — but it is showing you the wrong question.

Problem 1: Blended ROAS hides brand vs. non-brand

A Google Ads account with 8x blended ROAS often looks like this underneath:

  • Brand keywords (your own shop name): 40x ROAS
  • Non-brand keywords (generic product terms): 2x ROAS

Brand keywords convert customers who were going to buy from you anyway. You are paying Google to not lose them to a competitor bidding on your brand. That is fine — but it is not growth.

Non-brand at 2x ROAS is the real story. If your break-even is 3x (after product cost, shipping, returns), you are losing money on every non-brand click. Scaling non-brand actually destroys margin, even though your blended ROAS looks healthy.

Problem 2: Last-click attribution is nearly always wrong

Most platforms default to last-click. A customer sees 5 Meta ads, Googles your brand, then clicks the Google Brand ad and buys. Last-click gives 100% credit to Google Brand.

Meta shows its own attribution (often 7-day click + 1-day view), claiming credit for the same order. Google does the same. Sum it all up and your “total tracked revenue” from ad platforms exceeds your actual revenue by 30–80%.

Which one is right? None of them are, alone.

Problem 3: Marginal ROAS vs. average ROAS

Your overall Meta account is 4x ROAS. You decide to double your Meta budget. Does revenue double?

Almost never. The first €1,000 spent on Meta targets your best audiences. The second €1,000 targets slightly worse audiences. By the time you reach €10,000, you are showing ads to people who were never going to buy.

Marginal ROAS — the ROAS on the next euro spent — is almost always lower than average ROAS. If marginal ROAS drops below break-even, scaling is unprofitable even with a healthy average.

Problem 4: Incrementality is ignored

A customer adds a product to cart. You retarget them with a Meta ad. They click the ad and buy. Meta claims credit.

But would they have bought anyway? Probably yes. Most retargeting takes credit for purchases that would have happened without it.

True incrementality asks: how much revenue am I getting that I would not have gotten without this channel? For retargeting, often 20–40% of claimed revenue is incremental. The rest is self-attribution.

What to track instead

1. ROAS by brand vs. non-brand

Split your Google Ads account into brand-only and non-brand-only reports. Track them separately. You need both, but they tell completely different stories.

2. Marginal ROAS at current spend level

Before you increase budget on any channel, estimate marginal ROAS. Response curves from tools like BlueWalnut Genius show you the diminishing returns. If marginal ROAS is below break-even, do not scale — optimize targeting first.

3. CLV:CAC by acquisition channel

A channel with 2x ROAS on first purchase can be excellent if those customers return 3 more times. A channel with 5x ROAS on first purchase can be bad if they never come back. See our CLV guide.

4. Contribution margin ROAS (not revenue ROAS)

Revenue ROAS divides revenue by ad spend. Contribution margin ROAS divides gross profit by ad spend. A €100 sale at 20% margin contributes €20 to covering your overhead. If ad spend was €25, you lost money — even at 4x revenue ROAS.

5. Cross-channel attribution with Shapley values

Rather than assigning 100% credit to one touchpoint, Shapley-value attribution distributes credit across all channels based on their marginal contribution. Complex to calculate manually — but it is the only honest way to compare channels.

The blended ROAS myth

Many agencies report “total marketing ROAS” by dividing total revenue by total ad spend. This number is meaningless for decision-making because:

  • It includes organic/direct traffic revenue not driven by ads
  • It mixes brand and non-brand
  • It ignores channel-specific incrementality
  • It cannot guide budget decisions

A 6x blended ROAS tells you nothing about whether to increase Google spend or Meta spend or cut both.

What good ROAS analysis looks like

A useful ROAS dashboard shows, for each channel:

  • Average ROAS (revenue / spend)
  • Marginal ROAS at current spend
  • Response curve (how ROAS changes as spend scales)
  • Brand vs. non-brand split (for search channels)
  • First-order vs. CLV-adjusted ROAS
  • Estimated incrementality

With these six lenses, you can actually answer: “Should I scale this channel?” “Should I cut it?” “Where is the real growth coming from?”

Stop optimizing the wrong number

Most webshops improve their blended ROAS and wonder why profit does not grow. The reason: blended ROAS is not correlated with profitability at the margin.

BlueWalnut Genius breaks down ROAS by brand vs. non-brand, calculates marginal ROAS with response curves, and surfaces marginal waste automatically. You see which channels are genuinely scaling profitably — and which are just taking credit for organic demand.

Request a free demo — see your real channel-by-channel profitability.

Related articles