Why ROAS Is a Misleading Metric for Growing Brands
22 April 2026
ROAS is the most widely reported metric in performance marketing. It is also, for growing brands, one of the most misleading.
The problem is not that ROAS is wrong. It is that ROAS is narrow. It measures revenue returned per dollar of ad spend in a defined period. It does not measure profit, incrementality, the sustainability of the demand being converted, or whether the programme is building the conditions it will need to convert efficiently in the future. For a brand maintaining existing revenue, narrow is fine. For a brand trying to grow it, narrow is dangerous.
What ROAS Actually Fails to Measure
ROAS measures revenue efficiency. It does not measure profit, incrementality, demand generation, or long-term brand health.
These are not minor omissions. A campaign with 5x ROAS on a 10% margin product is losing money on every transaction. A campaign with 3x ROAS where 70% of conversions are organic traffic being attributed to paid channels is largely paying for conversions that would have happened without the ad. Neither version of this problem is visible in the ROAS figure. Both look fine in a dashboard.
The more fundamental limitation is time horizon. ROAS measures the return from this campaign, in this period, against this spend. It provides no signal about whether the programme is building the demand it will need to convert efficiently in the future. A growing brand needs both short-term conversion and long-term demand creation. ROAS observes only the first.
How Last-Click Attribution Inflates ROAS Figures
Last-click attribution assigns 100% of conversion credit to the final touchpoint before conversion.
In practice, paid search captures credit for conversions that began with a social ad, a brand campaign, organic content, or a word-of-mouth referral. The buyer's journey involved multiple brand interactions across days or weeks. The paid search ad was the final step. Last-click records it as the cause.
The result: paid search ROAS is systematically overstated. Upper-funnel channels are systematically understated. Budget flows toward the channel with the strongest attribution signal, not necessarily the channel doing the most work. Gartner found only 30% of CMOs are confident in their ability to measure marketing ROI accurately (Gartner, 2025: https://www.gartner.com/en/marketing/insights/articles/marketing-analytics-trends). Attribution error is a primary reason. ROAS looks authoritative while reflecting a model that systematically misallocates credit.
Even data-driven attribution, which distributes credit across touchpoints based on machine learning, cannot account for the brand exposure that happened before the tracked journey began: the billboard, the word-of-mouth recommendation, the category association built through two years of brand advertising. Those effects are real. They are entirely outside the attribution window. ROAS, regardless of the attribution model feeding it, measures only what the model can see.
Why a Rising ROAS Can Signal Brand Deterioration
A rising ROAS in a flat or declining revenue environment is a harvesting signal, not an efficiency gain.
When upper-funnel brand investment is cut, performance channels initially improve. The remaining audiences are more brand-aware and easier to convert. Conversion cost falls. ROAS rises. Every dashboard metric looks better.
What is actually happening: the programme is drawing down on brand equity built by the investment that was just cut. Warm audiences are being converted faster than new ones are being warmed. ROAS continues to rise as the algorithm finds the easiest remaining conversions. Revenue eventually plateaus as the depleted pool no longer provides enough volume. By the time the damage is visible in revenue, the connection to the brand spend reduction six to twelve months earlier is invisible to most reporting systems.
Analytic Partners found brand messaging outperforms performance messaging 80% of the time over the long term (Analytic Partners, 2022: https://www.marketingdive.com/). The harvesting dynamic that ROAS optimisation incentivises works directly against this. It rewards the programme for consuming the brand equity that generates long-term return, while making the consumption invisible in the metrics used to manage the programme.
What ROAS Misses About Upper-Funnel Activity
Upper-funnel activity creates the conditions that make lower-funnel performance possible. ROAS cannot see this relationship.
A buyer clicking a paid search ad is acting on a need they recognised, searching in language they used to describe it, and choosing from a consideration set shaped by previous brand exposure. The paid search ad is the final measurable step in a chain of interactions. ROAS credits only that final step.
Analytic Partners' analysis across thousands of marketing programmes found that brands combining brand and performance investment consistently outperform those running performance alone — not because brand campaigns convert directly, but because they increase the efficiency of every conversion channel they feed (Analytic Partners, 2022). ROAS cannot observe this effect because it measures within-channel return in isolation. The interdependency between brand and performance spend is real and measurable through media mix modelling. It is invisible to ROAS.
Which Metrics Should Sit Alongside ROAS in a Growth Dashboard
ROAS belongs in a dashboard. It should not be the sole metric driving strategic decisions. Five metrics together constitute a growth-oriented measurement stack, each measuring something ROAS cannot.
Contribution margin per acquisition measures actual profit per customer after all variable costs. ROAS measures revenue. A 4x ROAS at 15% margin is a loss. Contribution margin converts the revenue figure into a business outcome.
New customer acquisition rate tracks whether the channel is bringing in genuinely new buyers. ROAS counts conversions regardless of whether the customer is new or returning. A programme generating 3x ROAS with 70% returning customers is primarily re-acquiring existing buyers, not growing the business.
Branded versus non-branded traffic split uses branded search trend as a leading indicator of mental availability. ROAS is entirely agnostic to whether brand equity is growing or being depleted. A declining branded search trend while paid ROAS holds steady is a warning that the warm audience pool is shrinking.
Incremental ROAS measures return attributable to the campaign itself, not the organic baseline. Standard ROAS includes conversions that would have happened without the ad. Research suggests 30 to 50% of paid conversions in last-click models would have occurred organically (Nielsen, 2023: https://www.nielsen.com/). Incremental ROAS, derived from holdout or lift testing, shows the actual advertising effect.
LTV:CAC ratio measures the lifetime value of an acquired customer relative to acquisition cost. ROAS measures only the first transaction. For businesses with meaningful repeat purchase rates, LTV often exceeds first-transaction revenue by a multiple. Managing to first-transaction ROAS alone optimises for the wrong event.
How High-Growth Brands Use ROAS Without Being Misled by It
High-growth brands use ROAS as a constraint, not a target.
The constraint version sets a floor: the minimum acceptable return for a campaign or channel to remain funded, derived from margin analysis rather than from platform benchmarks or competitor comparisons. Any campaign clearing the floor is eligible for continued investment. The question then moves to volume and growth metrics, not further efficiency improvement.
This prevents the harvesting dynamic. A brand optimising for maximum ROAS will continuously identify the cheapest conversions — the warmest, most brand-aware audiences — progressively excluding harder-to-reach, less warm prospects. A brand holding ROAS at a floor while optimising for acquisition volume expands its pool instead of narrowing it. Some efficiency is traded for new customer growth that sustains the programme over time. That trade is the growth programme. Pure ROAS maximisation is the harvesting programme.
The distinction between the two is not visible in a ROAS chart. It is visible in new customer acquisition trends, branded search data, and revenue trajectory over a 12 to 24 month horizon.
The Metric That Measures What Matters
ROAS was introduced at the start of this article as narrow rather than wrong. The narrowness is the problem. A brand maintaining existing scale can manage with a narrow metric. A brand trying to grow needs to see the full picture: profit per acquisition, new customer growth, brand health, and incrementality alongside the revenue ratio that ROAS provides.
The brands that scale without stalling are not the ones with the highest ROAS. They are the ones that set a defensible ROAS floor and then optimise for the metrics that predict future performance: new customer acquisition, brand health, and the sustainability of the demand pool they are converting from. ROAS is one measure of whether the engine is running. It says nothing about whether the fuel is running out.
If you are reviewing your measurement stack or want help building a framework that uses ROAS appropriately alongside the metrics it cannot replace, Kaliber works with brands on exactly this. Start the conversation at kaliber.asia/contact.
Frequently Asked Questions
Why is ROAS considered misleading?
ROAS measures revenue returned per unit of ad spend but excludes profit margin, incrementality, new customer acquisition, and brand health. A high ROAS can coexist with loss-making campaigns (if margins are thin), declining new customer growth (if audiences are narrowing), and deteriorating brand equity (if the warm audience pool is being depleted without replenishment). ROAS reports accurately on what it measures. The problem is what it does not measure.
What is the difference between ROAS and incremental ROAS?
Standard ROAS divides total attributed revenue by total ad spend, including conversions that would have occurred without the ad. Incremental ROAS measures only the return attributable to the campaign itself — the conversions caused by advertising rather than organic behaviour being credited to paid channels. Research suggests 30 to 50% of conversions attributed to paid channels in last-click models would have occurred organically. Incremental ROAS is derived from holdout testing or geo-based lift experiments.
How should ROAS be used correctly?
ROAS is most useful as a floor constraint rather than an optimisation target. Set a minimum acceptable ROAS from margin analysis and hold to it, then optimise for volume and new customer acquisition above that floor. Avoid using ROAS as the sole bidding target, as automated systems set to maximise ROAS will progressively narrow toward the easiest, warmest conversions and exclude the new customer acquisition that growth requires.
What metrics should replace or complement ROAS?
No single metric replaces ROAS. The metrics that contextualise it are: contribution margin per acquisition (profit rather than revenue), new customer acquisition rate (growth rather than recycling), branded versus non-branded traffic split (brand health trend), incremental ROAS (advertising effect rather than organic attribution), and LTV:CAC ratio (long-term value rather than first-transaction return). Together these produce a measurement stack that ROAS alone cannot provide.
Can rising ROAS be a bad sign?
Yes. When ROAS rises while revenue growth slows or stalls, it typically indicates that the programme is narrowing toward warm existing audiences rather than reaching new ones. This is the harvesting dynamic: the algorithm finds progressively easier conversions by concentrating on the most brand-aware buyers, which produces a rising ROAS while quietly depleting the demand pool that sustains the programme. New customer acquisition rate and branded search trend are the signals that reveal this when ROAS cannot.
