Table of Contents
Blended ROAS is total revenue divided by total paid media spend across every channel, measured against the business as a whole rather than against any single ad platform. Where platform ROAS answers “what did this ad account claim it produced,” blended ROAS answers “what did the company actually earn for every euro it put into advertising.” The related metric, Marketing Efficiency Ratio (MER), widens the denominator further: MER is total revenue divided by total marketing cost, including ad spend plus agency fees, creative production, tooling, and influencer budgets.
The distinction stopped being academic the moment deterministic attribution broke. Admiral Media manages more than €500M in ad spend across 150+ mobile brands, and in the accounts the Admiral Media team runs today, the sum of platform-reported revenue routinely exceeds actual booked revenue. Every channel claims the same conversion. Blended ROAS is the arithmetic that cannot be double counted, which is precisely why it has become the number that decides budgets.
This guide covers what blended ROAS measures, how it differs from MER and platform ROAS, how to calculate both without corrupting the inputs, how to read the gap when platform and blended numbers disagree, and how Admiral Media sets defensible blended targets using real case study evidence.
What blended ROAS actually measures
Blended ROAS measures the efficiency of your entire paid media portfolio as a single unit, using revenue your finance system can verify rather than revenue your ad platforms estimate. The formula is deliberately blunt: total revenue in a period divided by total paid media spend in that same period.
The bluntness is the feature. Platform ROAS is computed inside an attribution model that each ad network controls, using click and view windows each network sets, on conversions each network claims. Blended ROAS is computed outside all of them. It cannot be inflated by a longer lookback window, it cannot be gamed by view-through credit, and it cannot double count a user who saw an ad on three networks before subscribing.
What blended ROAS gives up in exchange is diagnostic resolution. It tells Admiral Media clients whether the portfolio is healthy. It does not tell them which campaign to pause. That trade is the entire reason serious measurement stacks run platform ROAS, blended ROAS, and MER simultaneously rather than choosing one.
The three numbers and what each one is for
In Admiral Media’s reporting practice, each metric owns a different decision at a different cadence. Confusing them is the most common source of bad budget calls the Admiral Media team encounters when auditing new accounts.
| Metric | Formula | Denominator scope | Decision it owns | Review cadence | Primary failure mode |
|---|---|---|---|---|---|
| Platform ROAS | Platform-attributed revenue ÷ platform spend | One ad account | Which campaign, ad set, or creative to scale or kill | Daily to every 3 days | Double counting across networks; inflated by view-through credit |
| Blended ROAS | Total revenue ÷ total paid media spend | All paid channels | How much total budget the business can deploy | Weekly | Cannot isolate a bad channel; contaminated by organic revenue |
| MER | Total revenue ÷ total marketing cost | All paid channels plus agency, creative, tooling, influencer | Whether the marketing function as a whole is profitable | Monthly to quarterly | Too slow and too coarse to guide in-account optimization |
| Incremental ROAS | Lift-derived revenue ÷ spend on the tested channel | One channel, causally isolated | Whether a channel deserves to exist at all | Quarterly test cycles | Expensive; requires holdouts and statistical power |
Platform ROAS is the scalpel. Blended ROAS is the thermometer. MER is the balance sheet. Incrementality testing is the autopsy that tells you whether the patient needed the treatment. Admiral Media runs all four because each one is blind in a way the others are not.
Blended ROAS vs MER: the difference that matters
Blended ROAS and MER are not synonyms, and treating them as interchangeable is how marketing teams end up defending a number their CFO does not recognise. The difference is entirely in the denominator.
Blended ROAS counts media spend only. If a subscription app spends €400,000 across Google, Meta, TikTok, and Apple Search Ads in a month and books €1,200,000 in revenue, blended ROAS is 3.0. MER counts everything the marketing function consumes. Add a €40,000 agency retainer, €25,000 in creative production, €10,000 in measurement tooling, and €25,000 in creator fees, and total marketing cost becomes €500,000. MER is 2.4.
That gap between 3.0 and 2.4 is not noise. It is the true cost of running the growth engine, and it is the number that survives contact with a board deck. In Admiral Media’s experience, teams that report only blended ROAS systematically overstate efficiency by excluding the fixed costs that scale with headcount and creative volume rather than with media.
Which one to set targets against
Admiral Media sets in-account optimisation targets against platform ROAS, weekly deployment decisions against blended ROAS, and profitability commitments against MER. The reason is structural. Media spend is the only variable a media buyer can move today. Agency and tooling costs are step functions that change quarterly. Holding a buyer accountable to MER punishes them for costs they cannot influence inside a two-week window.
The corollary matters more. If MER is deteriorating while blended ROAS holds flat, the problem is not media efficiency, it is cost structure. Admiral Media has seen accounts where blended ROAS improved for two consecutive quarters while MER declined, because creative production volume and tooling licences grew faster than revenue. No amount of bid optimisation fixes that.
Why attribution decay made blended ROAS the default
Blended ROAS moved from a finance-side sanity check to the primary efficiency metric because platform-level attribution stopped being reliable enough to sum. Three structural changes drove this.
First, App Tracking Transparency removed deterministic device-level identifiers for the majority of iOS users, forcing networks onto probabilistic and modelled attribution. Second, SKAdNetwork replaced user-level reporting with delayed, coarse, privacy-thresholded postbacks, which means the revenue a network reports is a reconstruction rather than a record. Apple’s successor framework, AdAttributionKit, preserves the same privacy architecture: cryptographically signed postbacks with limited, tiered detail rather than raw user-level joins. Third, every major network now models unobserved conversions to fill the gap, and every network models them optimistically.
The arithmetic consequence is unavoidable. When Google, Meta, and TikTok each model the same install and each claim credit under their own attribution window, the sum of platform-reported revenue exceeds real revenue. Admiral Media routinely finds combined platform-attributed revenue substantially above verified booked revenue in accounts inherited from previous management. Blended ROAS is immune to this because its numerator comes from the payment processor or the app store, not from the ad account.
For teams still building their signal architecture, the encoding decisions in SKAdNetwork conversion values determine how much revenue fidelity a network ever sees. Poor conversion value design produces platform ROAS that is not just inflated but directionally wrong, which makes the blended check mandatory rather than optional.
The Admiral Media Blended Efficiency Ladder
Most teams adopt blended ROAS by simply dividing two numbers and then discovering the number is unusable, because organic revenue, refunds, and channel mix shifts contaminate it. Admiral Media built the following framework to move an account from a raw blended figure to a target that can actually drive budget decisions.
The Admiral Media Blended Efficiency Ladder
- Fix the revenue source of truth. Choose one system, typically the payment processor, the app store financial report, or the subscription management layer, and use it for every blended calculation. Net of refunds, chargebacks, and store commission. If revenue can be pulled from three systems that disagree, blended ROAS is meaningless before you start.
- Define the spend perimeter explicitly. Decide what counts as paid media and write it down. Ad platform spend is obvious. Affiliate commissions, app store search ads, and paid influencer flat fees are judgment calls. The perimeter matters less than its consistency across periods.
- Establish the organic baseline. Measure revenue during a period of minimal or zero paid activity, or use a geo holdout, to estimate what the business earns without media. Blended ROAS that includes an unmeasured organic base flatters every channel and hides deteriorating paid performance behind brand momentum.
- Align the measurement window to the revenue model. For subscription apps, same-day spend against same-day revenue is nonsense, because today’s spend produces revenue over the following twelve months. Use cohort-aligned windows or trailing-period matching that reflects when revenue actually lands.
- Set the target from unit economics, not from history. Derive the blended ROAS floor from contribution margin and payback tolerance, not from last quarter’s average. A blended ROAS of 2.0 is excellent for a business with 80% gross margin and catastrophic for one at 25%.
- Instrument the divergence, not just the level. Track the ratio of summed platform-reported revenue to verified blended revenue every week. The trend in that ratio is an early warning system: a widening gap means modelled attribution is drifting and in-account decisions are being made on increasingly fictional inputs.
- Validate causally at least once per quarter. Blended ROAS is correlational. Confirm with geo holdouts, PSA tests, or a marketing mix model that the channels absorbing budget are the ones producing the revenue.
Steps one through four are hygiene. Most accounts Admiral Media audits fail at step one or step three, which means every downstream decision was built on a contaminated denominator or an unmeasured organic base.
How to calculate blended ROAS and MER correctly
The formulas are trivial and the inputs are where the errors live. Blended ROAS equals total net revenue divided by total paid media spend for the same period. MER equals total net revenue divided by total marketing cost for the same period.
Getting the numerator right
Use net revenue, not gross bookings. For app businesses this means revenue after app store commission, after refunds, and after chargebacks. A 30% store commission applied to a gross figure inflates blended ROAS by roughly 43% relative to net, which is more than enough to justify a budget increase that the business cannot actually afford.
Decide once whether the numerator includes organic revenue. Both conventions are defensible. Including it produces a business-level efficiency read. Excluding it, by subtracting an estimated organic baseline, produces a paid-media efficiency read that is harder to compute but far more actionable. Admiral Media’s default is to report both, labelled clearly, because the two answer different questions.
Getting the denominator right
Include every paid media line item, including the ones that live outside the main ad accounts. App store search ads, programmatic DSP spend, retargeting platforms, and paid partnerships all belong in the paid media denominator for blended ROAS. Excluding a channel because it is small or because its reporting is inconvenient produces a blended figure that improves every time you add an unmeasured channel.
Getting the window right
This is where subscription apps break the metric. Spend on 1 August produces revenue in August, September, October, and beyond. Dividing August revenue by August spend during a scaling period understates efficiency, because a growing share of August spend has not yet returned. During a contraction it overstates efficiency, because you are harvesting revenue from cohorts acquired months earlier.
Admiral Media handles this two ways. For directional weekly reads, use trailing 28-day or trailing 90-day matched windows, which smooth the distortion without eliminating it. For decision-grade reads, switch to cohort economics and evaluate against CAC payback period rather than a period ratio. Blended ROAS on a trailing window tells you the direction. Cohort payback tells you whether the business can fund the growth.
Reading the gap: when platform and blended ROAS disagree
The gap between summed platform ROAS and blended ROAS is diagnostic information, not an error to be reconciled away. Admiral Media treats the direction and the trend of that gap as a leading indicator of measurement health.
| Platform ROAS | Blended ROAS | Most likely cause | Recommended action |
|---|---|---|---|
| Rising | Flat | Channels are cannibalising each other or claiming credit for the same conversions | Run a geo holdout on the channel claiming the largest incremental gain |
| Rising | Falling | Modelled attribution is drifting upward while real incremental revenue declines | Freeze budget increases; validate with an incrementality test before scaling further |
| Flat | Rising | Organic or brand-driven revenue is growing, or a lagging cohort is maturing | Re-baseline organic; confirm the lift is not being wrongly credited to paid |
| Falling | Flat | Attribution loss, typically from signal degradation rather than performance decline | Audit conversion value encoding and consent rates before cutting spend |
| Falling | Falling | Genuine performance deterioration, usually creative fatigue or auction pressure | Refresh creative; check creative fatigue indicators before touching bids |
The row that costs the most money is the second one. Platform ROAS rising while blended ROAS falls is the signature of an account being scaled on modelled revenue that is not landing in the bank. It is also the single most common pattern Admiral Media finds when taking over accounts that were scaled aggressively on in-platform numbers alone.
Case evidence: what channel divergence looks like in real accounts
Blended ROAS is a portfolio number, but portfolios are built from channels that behave differently. Two Admiral Media engagements show why a single blended target has to be decomposed before it can be acted on.
ChatPDF: two channels, two different efficiency curves
Admiral Media restructured ChatPDF’s paid acquisition across Google and Meta, consolidating overlapping ad sets, moving from target CPA to value-based bidding with LTV signal, and running a cadenced weekly creative testing programme. In Admiral Media’s work with ChatPDF, the campaign delivered +320% ROAS, +156% subscriptions, and -42% CAC overall, measured on an index baseline comparing Year 1 against Year 2 year to date.
The channel split is the interesting part. Google produced 320% ROAS year-on-year growth, +142% subscription growth, and a 38% CAC reduction. Meta produced 280% ROAS year-on-year growth, +171% subscription growth, and a 45% CAC reduction. The channel that won on ROAS growth was not the channel that won on volume growth or on cost reduction.
A single blended ROAS target would have obscured this entirely. The portfolio number would have looked healthy while hiding the fact that the two channels were optimising toward different outcomes. This is exactly why Admiral Media pairs a blended target with channel-level decomposition rather than replacing one with the other.
PURE: channel selection moved efficiency more than bid tuning
For the dating app PURE, Admiral Media tested a programmatic DSP against an established self-attributing network on US Android, allocating distinct budgets to each and tailoring creative to the requirements of each platform. In Admiral Media’s work with PURE, the programmatic route delivered a CPI of $2.44 against the self-attributing network’s $9.43, roughly four times lower, alongside a 74% CPI reduction and D7 ROAS goals that were exceeded, which unlocked new market launches.
The lesson for blended measurement is direct. When one route acquires users at roughly a quarter of the cost of another, portfolio blended ROAS is dominated by mix rather than by bid strategy. Shifting budget between channels moved efficiency more than any in-account bid adjustment could have. A blended target with no mix decomposition would have registered the improvement without ever explaining it.
Setting a blended ROAS target you can defend
A defensible blended ROAS target is derived from contribution margin and payback tolerance, never inherited from a benchmark. The mechanism is straightforward: your blended ROAS floor is the point at which revenue covers variable cost of delivery plus media, over the horizon your cash position can tolerate.
Three inputs determine it. Gross margin after store commission and delivery cost sets the ceiling on how much of each revenue euro can fund acquisition. Payback tolerance, expressed in months, sets how long the business can wait. Retention shape determines how much of the eventual LTV lands inside that window. An app with 80% margin and a twelve-month payback tolerance can run a same-month blended ROAS well below 1.0 and still be highly profitable. An app with thin margins and a three-month runway cannot.
This is why Admiral Media pairs blended ROAS with cohort-based measures rather than treating it as a standalone target. Predictive LTV bidding feeds value signal into the platforms so that in-account optimisation pulls toward the same outcome the blended target is measuring, and retention-first acquisition ensures the cohorts arriving actually mature into the revenue the target assumes.
Full-funnel improvement is what moves the blended number
Blended ROAS improves when the whole acquisition chain improves, not when a single lever is pulled. Admiral Media’s fifteen-month engagement with NeuroNation, a German brain training company, illustrates the pattern. The Admiral Media team ran intensive creative testing and channel exploration, categorised communication ideas and tested them against target audiences across all target markets, and introduced the pRank methodology to identify winning and losing variants faster.
In Admiral Media’s work with NeuroNation, the programme delivered +117% ROAS, +66% installs, +32% purchases, +42% net cohort revenue, and a 39% reduction in CPI, using data from January to August 2019.
Note the shape of the result. Installs grew faster than purchases, CPI fell, and net cohort revenue grew 42% while ROAS grew 117%. That combination is only possible when cost efficiency and revenue quality improve together. A team optimising against a blended target alone would have seen the ROAS number move without understanding which of the four underlying levers produced it.
Blended ROAS for subscription apps
Subscription apps need a modified blended approach because revenue arrives on a schedule that has nothing to do with when spend occurs. A monthly subscriber acquired today contributes revenue for as many months as they retain, which means period-matched blended ROAS is structurally misleading during any change in spend level.
Admiral Media applies three adjustments. First, report blended ROAS on a trailing 90-day window rather than a calendar month, which reduces the distortion from spend acceleration. Second, run a parallel cohort view that tracks D30, D90, and D180 revenue per acquisition cohort against the spend that produced it. Third, treat the period blended figure as a trend indicator and the cohort figure as the decision input.
Web-to-app flows complicate this further, because revenue can be recognised on the web layer while the acquisition cost sits in an app install campaign, or the reverse. Admiral Media’s guidance on web-to-app funnels covers the attribution plumbing, but the blended principle is simple: if a purchase path exists, its revenue belongs in the numerator and its acquisition cost belongs in the denominator, regardless of which surface the transaction closed on.
The organic contamination problem
Subscription apps with strong brand or app store presence face the largest organic contamination risk. If 40% of revenue arrives organically and that share is not held constant, blended ROAS moves for reasons that have nothing to do with media performance. Admiral Media establishes an organic baseline through geo holdouts or spend-down periods, then reports paid-adjusted blended ROAS alongside the raw figure so that neither number is used in isolation.
Common mistakes that corrupt blended measurement
Most blended ROAS programmes fail on data hygiene rather than on methodology. These are the errors Admiral Media encounters most frequently in account audits.
- Mixing gross and net revenue: Using gross bookings in the numerator while comparing against net-based targets systematically overstates efficiency by the size of the store commission.
- Excluding inconvenient spend: Leaving app store search ads, affiliate commissions, or a small test channel out of the denominator makes blended ROAS improve every time a new channel is added.
- Same-period matching for delayed revenue: Dividing this month’s revenue by this month’s spend in a subscription business measures the pace of scaling, not efficiency.
- Ignoring the organic baseline: Blended ROAS that includes an unmeasured organic base hides paid deterioration behind brand momentum until the trend is too far gone to reverse cheaply.
- Using blended ROAS for in-account decisions: A portfolio number cannot tell you which ad set is failing. Teams that pause campaigns based on a blended figure remove the wrong campaigns.
- Changing the definition mid-year: Any change to the spend perimeter or revenue source breaks period comparability. If the definition must change, restate history alongside it.
- Treating blended ROAS as causal: Blended ROAS is a correlation between two aggregates. Only incrementality testing or a properly specified mix model establishes causation.
How Admiral Media operationalises blended reporting
A blended ROAS programme is a reporting cadence and a set of ownership rules, not a dashboard. Admiral Media structures it so that each metric drives decisions at the frequency it can actually support.
| Cadence | Primary metric | Supporting checks | Decision authority | Typical action |
|---|---|---|---|---|
| Daily | Platform ROAS and CPA by campaign | Spend pacing, learning phase status, creative delivery share | Media buyer | Pause, scale, or rotate creative within the account |
| Weekly | Blended ROAS, trailing 28 days | Platform-to-blended revenue gap; channel mix shift | Account lead | Reallocate budget across channels; set next week’s spend level |
| Monthly | MER and cohort payback | D30 and D90 cohort revenue; organic baseline check | Growth lead and finance | Adjust total marketing budget; revisit target floors |
| Quarterly | Incremental ROAS by channel | Geo holdout or PSA test results; mix model refresh | Leadership | Add, cut, or restructure channels; reset annual targets |
The rule that makes this work is that no metric escalates a decision above its cadence. A single bad blended week does not trigger a channel cut, because blended ROAS at that resolution is too noisy to support the decision. A single bad campaign day does not trigger a budget reduction, because platform ROAS at that resolution says nothing about the portfolio.
Bidding alignment
Blended targets only work if in-account bidding pulls in the same direction. Value-based bidding is the mechanism, and it has a volume prerequisite. Google’s documentation for Target ROAS bidding states that the strategy should have at least 15 conversions in the last 30 days at the conversion tracking level, and recommends evaluating performance over periods containing at least 50 conversions. Accounts below that threshold will produce erratic tROAS behaviour regardless of how well the blended target is specified, which is why Admiral Media often runs volume-based bidding with proxy events first and migrates to value bidding once the conversion base supports it. Google’s broader Smart Bidding documentation sets out the signal inputs these strategies use.
For a fuller view of which metrics deserve a place in the reporting stack and which are vanity, see Admiral Media’s guide to app marketing metrics that actually matter.
Frequently Asked Questions
What is a good blended ROAS for a mobile app?
There is no universal benchmark, because the answer depends entirely on gross margin and payback tolerance. A blended ROAS of 2.0 can be highly profitable for a subscription app with 80% margin after store commission and a twelve-month payback horizon, and unsustainable for a business with 25% contribution margin and a three-month cash runway. The correct approach is to derive the floor from your own unit economics: calculate the contribution margin per revenue unit, decide how many months of payback the business can fund, then solve for the blended ROAS that satisfies both. Any benchmark quoted without those two inputs is not usable.
What is the difference between blended ROAS and MER?
Blended ROAS is total revenue divided by total paid media spend. MER, or Marketing Efficiency Ratio, is total revenue divided by total marketing cost, which includes media spend plus agency fees, creative production, measurement tooling, and influencer or creator payments. Blended ROAS therefore always reads higher than MER for the same period. Blended ROAS is the right metric for weekly media budget decisions because media spend is the variable a buyer can actually move. MER is the right metric for monthly and quarterly profitability discussions because it captures the full cost of running the marketing function.
Why is my platform ROAS higher than my blended ROAS?
Almost always because multiple ad networks are claiming credit for the same conversions. Since App Tracking Transparency and SKAdNetwork removed deterministic device-level attribution, each network models unobserved conversions using its own methodology and its own attribution window, and each does so optimistically. When those independently modelled figures are summed, the total exceeds real revenue. Blended ROAS uses verified revenue from your payment processor or app store report, so it cannot double count. A persistent gap is normal; a widening gap is a signal that modelled attribution is drifting and that in-account decisions are increasingly being made on unreliable inputs.
Should blended ROAS include organic revenue?
Both conventions are defensible, but they answer different questions and should not be mixed within a single report. Including organic revenue produces a business-level efficiency read that is easy to compute and useful for leadership. Excluding it, by subtracting an estimated organic baseline established through a geo holdout or a spend-down period, produces a paid-media efficiency read that is far more actionable for the media team. Admiral Media’s default is to report both figures side by side with explicit labels, because a blended number that silently includes a growing organic base will hide deteriorating paid performance for months.
How often should blended ROAS be reviewed?
Weekly, on a trailing 28-day window, is the practical cadence for most app businesses. Daily blended ROAS is too noisy to support decisions, particularly for subscription apps where revenue lands on a delay. Monthly is too slow to catch a deteriorating trend before meaningful budget has been misallocated. The critical discipline is that the review cadence must match the decision the metric owns: blended ROAS governs weekly budget allocation across channels, while platform ROAS governs daily in-account optimisation and MER governs monthly profitability decisions.
Can blended ROAS replace attribution entirely?
No. Blended ROAS tells you whether the portfolio is efficient, but it cannot tell you which channel, campaign, or creative produced the result, which means it cannot guide optimisation. It is also correlational rather than causal, so a rising blended ROAS during a period of brand growth or seasonal demand can mask genuinely unprofitable media. A complete measurement stack pairs blended ROAS with platform-level attribution for tactical decisions, incrementality testing for causal validation, and marketing mix modelling for long-horizon budget allocation. Each method is blind in a way the others are not.
How do subscription apps calculate blended ROAS without distorting it?
Use a trailing 90-day window for the period figure and run a parallel cohort view as the decision input. Period-matched blended ROAS is structurally misleading for subscription businesses because spend today produces revenue over the following months, so the metric understates efficiency while scaling and overstates it while contracting. The cohort view tracks D30, D90, and D180 revenue for each acquisition cohort against the spend that produced it, which is the only view that answers whether acquired users pay back. Evaluate against CAC payback period rather than a single-period ratio when making budget commitments.
