Blended vs Per-Channel ROAS: Which Should You Trust?

DTC brands get into trouble when they optimize per-channel ROAS in isolation. Google looks great. Meta looks weak. The obvious move is to cut Meta and scale Google. Six months later, Google performance collapses because the top-of-funnel awareness that Meta was building has dried up.

This is the channel attribution trap, and the way out is understanding what blended ROAS (also called MER) and per-channel ROAS are each actually measuring.

What per-channel ROAS tells you

Per-channel ROAS measures the revenue each platform claims credit for, divided by what you spent on that platform. Meta ROAS = revenue from Meta-attributed orders / Meta spend. Google ROAS = revenue from Google-attributed orders / Google spend.

The limitation is that attribution is imperfect. Both platforms measure within their own attribution windows using their own logic. Meta's 7-day click + 1-day view attribution means it claims credit for purchases that happened up to a week after someone saw an ad. Google uses last-click by default and claims the conversion when someone searched for your brand and clicked through.

In practice, both platforms often claim credit for the same order. If you add up the attributed revenue from all channels, it frequently exceeds your actual Shopify revenue. This is normal, and it is why per-channel ROAS alone is misleading.

What blended ROAS (MER) tells you

MER is simple: total Shopify revenue divided by total ad spend across all channels. It uses actual revenue from your order management system, not attributed revenue from any ad platform.

MER = Total Revenue / Total Ad Spend (all channels)

MER has no attribution problem. It does not matter which platform claims credit. You look at what you spent on ads in total, and what revenue the business generated. If MER is above your break-even ROAS, the overall marketing engine is profitable. If it is below, the whole business is losing money on advertising regardless of what the per-channel ROAS numbers say.

The combined view: how to use both

• Use MER as your primary health metric: check it daily. If MER drops below your break-even ROAS, something is wrong at the whole-business level and you need to investigate.
• Use per-channel ROAS to allocate budget: if Meta ROAS drops significantly while Google ROAS holds, that suggests a Meta-specific problem (creative fatigue, audience saturation, CPM spike) that needs attention.
• Do not scale Google by cutting Meta without watching MER: the most common mistake. If cutting Meta spend causes MER to drop even though Google ROAS looks fine, Meta was providing value that Google could not show in its own metrics.
• Run incrementality tests: the most reliable way to understand channel contribution is to temporarily reduce or pause spend on a channel and watch what happens to total revenue. This is the only clean way to isolate true lift.

Signals that per-channel ROAS is misleading you

• Attributed revenue from all channels combined is significantly higher than Shopify revenue (more than 1.3-1.5x is a red flag).
• Google brand search ROAS looks excellent but Google non-brand ROAS is weak: you may be paying to capture demand that Meta and TikTok are generating.
• Meta ROAS has been declining for months but MER is stable: Meta may be doing top-of-funnel work that Google is closing. Cutting Meta would likely hurt MER.
• MER is flat or rising but per-channel ROAS across all channels is declining: new-customer ROAS may be declining while returning customer purchases inflate blended per-channel numbers.

The short answer

Trust MER for business-level decisions. Trust per-channel ROAS for channel-level decisions. They answer different questions. MER tells you whether the whole marketing engine is healthy. Per-channel ROAS tells you where to reallocate within that engine.

Set a MER floor based on your contribution margin (your break-even ROAS). Track it daily. Within that constraint, use per-channel data to optimize allocation. This is the framework that prevents both overspending on a broken channel and making the mistake of cutting channels that are doing invisible top-of-funnel work.