Why it matters
Each ad platform over-attributes in its own dashboard. Meta, Google, and TikTok can each claim credit for the same customer journey. Summing platform-reported revenue without deduplication routinely inflates numerators. Blended ROAS forces a single attribution story and one revenue definition across spend.
Blended ROAS is the metric growth and finance often use for weekly business reviews: total attributed revenue divided by total ad spend. It is useful for pacing and mix decisions, but it still mixes correlation with causation unless paired with incrementality work.
For LTV-driven businesses, blended ROAS on first-purchase revenue can look healthy while cohort profit weakens after returns and low repeat. Teams moving to value-based bidding on pLTV need blended readout at maturity to see whether cross-channel efficiency improved, not only whether one platform's dashboard rose.
Blended ROAS
Blended ROAS is a common executive readout for pLTV pilots, but only if numerators and windows align with the value definition you activated:
- First-party data in your data warehouse supplies orders, subscriptions, returns, and customer IDs for deduplicated revenue.
- Churney models pLTV and sends values directly to ad networks via Meta CAPI and the Google Ads Conversion API for treatment campaigns.
- During the pilot, compare blended ROAS (internal attribution) against platform ROAS per network to spot overlap and signal delivery issues.
- At cohort maturity, recompute blended ROAS on realized revenue (net of returns where relevant), not only on sent predicted values.
- Pair blended efficiency with incremental ROAS from a holdout test before declaring the signal a win.
Blended ROAS summarizes portfolio efficiency. It does not replace experiment readout when you need to know if pLTV caused lift.
Generic form:
Blended ROAS = Blended attributed revenue / Total ad spendBuild checklist:
Spend: Sum platform invoices or API spend exports; exclude non-media fees if your definition requires it.
Revenue: Join conversions to customer and order data in your data warehouse; apply one attribution model.
Deduplication: One primary conversion per customer journey window where possible.
Maturity: For LTV businesses, restate at cohort maturity when repeat and refunds matter.
Category variants
| Model | How blended ROAS shows up |
|---|---|
| Ecommerce / DTC | Shopify or data warehouse revenue vs total Meta + Google + TikTok spend; returns often require net revenue numerator for honest readout. |
| Subscription app | Blended spend across UA channels vs realized subscription revenue; trial-start attribution may overstate early blended ROAS. |
| SaaS / PLG | Multi-touch or last-click internal model vs paid spend; expansion revenue may arrive outside default blended windows. |
Common mistakes
- Summing platform numerators. Double-counting conversions across networks inflates blended ROAS without a deduplication layer.
- Mixing gross revenue with net margin goals. Blended ROAS looks strong until refunds and COGS are applied at maturity.
- Changing attribution rules mid-quarter. Model or window shifts make trend lines non-comparable.
- Using blended ROAS as incremental proof. Portfolio efficiency can move for reasons unrelated to a pLTV or bidding change.
- Ignoring organic and brand lift. Blended attribution may still credit paid touchpoints for demand that would have arrived anyway.
- Short windows on repeat-heavy categories. Blended ROAS peaks on first order before repurchase and return curves stabilize.
Advertiser lens
| Role | What they ask | What good looks like |
|---|---|---|
| Head of Performance / UA | What is our true cross-channel ROAS? | Documented attribution model, deduplication method, and revenue definition stable quarter to quarter. |
| VP Growth / CMO | Can we increase spend and stay efficient? | Blended ROAS trend with volume, channel mix, and incremental tests on major changes. |
| Marketing Analytics / Data Science | How do we build blended ROAS? | Single customer ID graph, agreed window, sensitivity checks vs platform totals. |
| Data Engineering | Can we join ad spend to revenue reliably? | Consistent UTM and click ID capture; data warehouse tables for spend and outcomes. |
| Finance / Procurement | Does blended ROAS tie to P&L? | Numerator mapped to recognized revenue or contribution margin; clear lag assumptions. |
FAQ
What is blended ROAS?
Blended ROAS is total attributed revenue (under your cross-channel attribution rules) divided by total ad spend across the channels included in the calculation.
How is blended ROAS different from platform ROAS?
Platform ROAS uses one ad network's attributed revenue and spend. Blended ROAS uses your internal or multi-channel attribution model and typically deduplicates credit across platforms.
How do you calculate blended ROAS?
Blended ROAS equals blended attributed revenue divided by total ad spend. The hard part is defining attributed revenue consistently: pick model (last click, MTA, etc.), window, and revenue basis (gross vs net).
Why is blended ROAS lower than platform ROAS?
Deduplication, stricter windows, and net revenue numerators usually reduce credited revenue versus summing platform dashboards.
Can blended ROAS prove a pLTV pilot worked?
It shows portfolio efficiency movement but not causation. Use holdout tests or structured experiments for incremental ROAS before scaling on signal changes.
Should blended ROAS use purchase value or pLTV?
For readout at maturity, use realized revenue (or agreed net revenue). pLTV is what platforms learn on during the test; blended ROAS readout should reflect outcomes, not predictions alone.
When should finance trust blended ROAS?
When attribution rules, revenue definitions, and spend sources are documented, stable, and reconciled to internal systems at agreed lag.
Not the same as
| Term | Difference |
|---|---|
| Platform ROAS | Single-network attribution; no cross-channel deduplication. |
| Incremental ROAS | Measures lift from a change vs control; blended ROAS is observational portfolio efficiency. |
| Media mix modeling (MMM) | Statistical channel contribution model; different method, often longer lag, not daily blended ROAS. |
| LTV reporting | Customer-level lifetime economics; blended ROAS is spend-period efficiency, not cohort LTV alone. |
