Return on Ad Spend (ROAS)
Return on ad spend (ROAS) is the revenue generated by an ad campaign divided by what the campaign cost, usually shown as a ratio like 4x or a percentage like 400%.
Why it matters
It connects advertising to money rather than to activity, which is the only comparison that actually settles a budget argument. Clicks, impressions and even leads can all rise while a business loses money; ROAS is the metric that forces the conversation back to revenue.
What people get wrong
Reporting platform ROAS as though it were business ROAS. The platform only knows what it recorded, not what you invoiced, and for service businesses with offline closes the two can differ enormously. It is also a revenue figure, not a profit figure - a 4x return on a job with a 20% margin is a loss once costs are counted.
How the number actually gets built
ROAS is a division problem sitting on top of two much messier inputs: the cost figure and the revenue figure. Cost is usually solid - the platform knows exactly what it spent. Revenue is not. It depends on a conversion value being attached to each recorded conversion, and that value has to come from somewhere: a fixed number set once when the conversion action was created, a dynamic value pulled from a form field or CRM, or an estimated value assigned after the fact.
For an ecommerce account the value is usually the actual order total, passed automatically at checkout, so platform ROAS and real ROAS tend to track closely. For a home services account there is rarely a checkout. The value attached to a 'booked appointment' conversion is frequently a flat placeholder - an average job size entered once and never revisited - so the ratio the dashboard reports is really 'revenue if every lead were an average job,' not revenue that occurred.
Smart Bidding strategies built around a target ROAS use this same value to decide which auctions to enter. If the value feeding the model is wrong, the algorithm is optimizing toward a number that does not describe your business, and it will do that very efficiently.
Where ROAS quietly goes wrong in real accounts
The most common failure is stale conversion values. A business raises its average job price, or starts selling a higher-margin service line, and nobody updates the value on the conversion action. The dashboard keeps reporting the old ratio, and a campaign that has actually become more profitable looks flat or declining, because the revenue side of the equation never moved with the business.
The second is counting every conversion at the same value regardless of what it actually was. A same-day repair and a multi-week installation are not the same job, but if both fire the same generic 'lead' conversion with the same flat value, the ROAS figure is an average that describes neither of them accurately - and Smart Bidding cannot tell them apart either.
The third, and the one that causes the most arguments, is comparing ROAS across a change in the attribution window or a change in which conversion actions are marked primary. The ratio can shift substantially with no change in actual sales, because the denominator or numerator moved, not the business. The symptom before anyone notices is usually a sudden step change in reported ROAS with no matching change in phone volume or job count - a sign to check the settings before questioning the media.
How ROAS relates to the rest of the account
ROAS is downstream of everything else on this list. It cannot be more accurate than the conversion tracking feeding it, and it inherits every quirk of the attribution window the account is using, which is why a ROAS figure for the most recent week is always the least reliable one - conversions are still arriving for it.
It is also a mirror image of cost per acquisition: CPA asks what one job cost, ROAS asks what the whole account returned relative to spend. Accounts that only watch one of the two tend to miss problems the other would have caught - a healthy ROAS can hide a rising CPA on a shrinking number of larger jobs, and a healthy CPA can hide a collapse in the value of each job.
Because ROAS is only ever as good as the value attached to a conversion, it is worth revisiting alongside conversion action settings whenever pricing, service mix, or average job size changes - not just when performance looks strange.
Frequently asked
Is a higher ROAS always better?
Not automatically. A very high ROAS on a campaign with a small number of conversions can just mean it is targeting a narrow, already-warm audience rather than growing the business. It is worth checking volume alongside the ratio - a lower ROAS at meaningfully higher volume can be the better outcome for overall revenue.
Why does my Google Ads ROAS not match my actual profit?
ROAS is a revenue figure, not a profit figure, and it usually reflects the value attached to a conversion action rather than an audited sale. Margin, refunds, cancellations and any offline adjustments never make it back into the platform, so the two numbers are answering different questions even when both are calculated correctly.
How often should conversion values be updated for ROAS to stay accurate?
Whenever pricing, service mix, or average job size changes meaningfully. A value set once at account setup and never revisited is one of the most common reasons ROAS drifts away from what the business is actually experiencing, in either direction.