Flight log / Transmission 064

ROAS for B2B: why it lies and what to use instead

In B2B, ROAS is a number from the future: the revenue shows up months after the click, and nobody has a DeLorean. Even in Google's own simple example, $1,200 in sales on $200 of ads is a 600% ROAS and a 50% ROI.

ROAS is the metric ecommerce runs on, and for good reason: a sale happens minutes after the click, the order value is known to the cent, and the margin is in the product catalog. B2B lead gen has none of that. The value is an estimate, the revenue lands months later, and the "conversion" is a form fill that might be a buyer or a student.

ROAS still shows up in B2B reports, often as the biggest number on the slide. This post covers what it actually measures, how it differs from ROI, the three ways it misleads B2B teams, and what to report instead.

What is ROAS?

ROAS, return on ad spend, is the value your ads produced divided by what they cost. In Google Ads the column is Conv. value / cost, which Google's help page on conversion tracking data describes as your total conversion value divided by the total cost of all ad interactions.

Google's Target ROAS help page uses a simple example: if your goal is $5 of sales for each $1 you spend on ads, that's $5 / $1 x 100% = a 500% target ROAS. Google expresses targets as percentages; many teams say "5x." Same number.

Everything depends on the top half of that fraction. Conversion value is whatever you tell Google a conversion is worth. Google's About conversion values page lists the options: the same value for every conversion, which is easier to set up but less representative, or different values per conversion, including transaction-specific values passed in from your site or CRM. In ecommerce that's the order total. In B2B lead gen it's usually a number someone picked.

What is the difference between ROAS and ROI?

ROAS divides revenue by ad cost. ROI subtracts all the costs of earning that revenue, then divides by those costs. Google's own ROI help page gives a clean example: a product costs $100 to make and sells for $200. You sell 6 through Google Ads, for $1,200 in sales, and your ads cost $200. Google calculates ROI as ($1,200 minus $600 minus $200) / ($600 + $200) = 50%.

Run the same numbers as ROAS and you get $1,200 / $200 = 600%. Same campaign, same month.

The last row is the one most reports skip. At a 50% gross margin, any ROAS under 200% loses money before you count salaries. At an 80% margin, break-even ROAS is 125%. A "good ROAS" only means something next to your margin.

One note on wording. Google's definition page for conversion value per cost says the metric "estimates your return on investment." It estimates one input to ROI. It doesn't account for cost of goods, sales costs or whether the revenue would have arrived anyway.

Why does ROAS mislead B2B teams?

Three reasons, each one fixable, none of them visible in the ROAS column itself.

1. The value is whatever you chose

Take one hypothetical B2B campaign that spends $20,000 in a month and produces 200 leads, 20 qualified opportunities and 5 closed deals worth $30,000 each in first-year revenue. Here's its ROAS under four reasonable value settings:

In this illustrative case, the same spend shows a 30x ROAS if you count pipeline ($600,000 across 20 opportunities), 7.5x on closed revenue ($150,000), 5.6x on gross profit and 1.0x if every lead is worth a flat $100. None of these is wrong. They answer different questions, and a report that just says "ROAS: 30x" hides which one it's answering. Pipeline value is the most common one in B2B decks, and it's the most generous, because pipeline isn't revenue.

2. The revenue arrives late, and some of it never arrives in Google Ads

Google's conversion tracking help says standard conversion columns are calculated based on the time of the click, not the time of the conversion. If you import closed deals, a deal that closes in March from a December click adds value to December, in March. So the most recent months always look weak, and they recover as deals close.

There's also a hard edge. Google's offline import guidelines say conversions uploaded more than 90 days after the last click won't be imported, and the limit is 63 days for enhanced conversions for leads. If your sales cycle is longer than that, part of your revenue will never appear in Google Ads ROAS at all.

In this illustrative account, the last three click months appear to collapse from about 7x to 0.6x. They haven't. Months 6 to 8 are still inside the 90-day import window, and their deals are still closing. A team that cuts budget on that chart is cutting the campaigns that are about to look best. And even the "settled" line is an undercount for any deal that took longer than 90 days.

3. ROAS can't tell you what was incremental

Platform ROAS credits the ad for every conversion it's attributed, including people who would have bought anyway. Brand search and remarketing are the usual offenders: they reach people already close to buying, so their ROAS looks spectacular. A holdout test, where you pause or withhold ads for a comparable group and compare results, estimates how much of that value the ads actually caused. Here's what that comparison can look like in a hypothetical account:

In this hypothetical, brand reports 18x but only 3x is incremental, while non-brand reports 4x and keeps 3.4x. Ranked on platform ROAS, brand gets the budget. Ranked on incremental ROAS, non-brand does. That's why budget shouldn't move between campaign types on platform ROAS alone. The incrementality testing post covers how to run a holdout without a data science team.

What should B2B teams use instead of ROAS?

Keep ROAS where it does a job, inside value-based bidding, and report decisions on metrics that hold up to a CFO's questions.

Cost per qualified opportunity is the workhorse. It doesn't depend on a value guess, it's available within weeks, and it's hard to inflate. Its weakness is that it treats a $10,000 deal and a $100,000 deal the same, which is where pipeline per dollar earns its place, as long as it's labeled honestly. Our cost per lead guide shows how to sanity check the earlier funnel stages, and the PPC report template shows how to fit all of this on one page.

Should B2B campaigns use Target ROAS bidding?

Sometimes. Target ROAS can be the right bid strategy for B2B when values reflect real differences, for example a demo request from a 500-person company worth more than one from a 5-person company. It's the wrong strategy when every lead carries the same static value, because then it's just Target CPA with extra steps.

Google's requirements and guidance, as of November 2026:

• Volume: Target ROAS on Search and Shopping needs at least 15 conversions in the past 30 days, per Google's Target ROAS help page.
• Values that differ: Google's conversion value best practices recommend measuring at least 2 unique values across at least 2 different actions, with neither value set to zero.
• The right funnel stage: the same page recommends optimizing to a stage with a relatively short conversion delay and at least 15 monthly conversions. In B2B that's often a qualified lead, not a closed deal.
• Set targets from history. Google suggests multiplying your historical conversion value per cost by 100 to get a starting Target ROAS percent, and warns that the target you set may influence the total conversion value you get. Set it too high and volume drops.

The value-based bidding guide covers how to set stage values that won't fool the algorithm.

Where this advice breaks down

• Self-serve B2B with card payments. If customers buy online at a known price, ROAS works much like ecommerce. Use it, with margin.
• No CRM data. Without imported stages, every alternative above is a manual spreadsheet. ROAS on a static value may be the best signal you have for now. Label it as such.
• Pipeline values drift. Sales teams revise deal sizes. If pipeline per dollar is on the report, fix the value at the stage it was created.
• Incrementality tests are expensive. Holdouts cost volume and take weeks. Run them on the biggest budget questions, not every campaign.

FAQ