ROAS (Return on Ad Spend)
ROAS doesn't measure whether the campaign works. It measures only whether it generates immediate revenue, and in B2B immediate revenue is almost always the smallest part of the real value.
“Measuring ROAS on last click, in a nine-month sales cycle with twelve touchpoints, is like giving all the credit for the goal to the final pass.”
Otto GTM ObservatoryThe structural limit of ROAS in B2B
ROAS was born in immediate-purchase contexts: a click, a cart, a conversion. In B2B, where an average deal takes weeks or months and involves multiple stakeholders and channels, attributing conversion credit to a single campaign systematically distorts reality.
A multi-touch attribution model (linear, time-decay, or algorithmic) returns a much more faithful picture of which touchpoints truly contribute along the journey, even when they aren't the last one before signature.
Anti-patterns
Common mistake: cutting budget for top-of-funnel awareness campaigns because they show low last-click ROAS, without considering they generated the first touchpoint of deals closed months later through other channels.
Second mistake: applying the same ROAS optimization logic used for e-commerce performance campaigns to enterprise account-based campaigns, where the cycle and number of stakeholders make the metric nearly inapplicable as-is.
Practical Application
An €18,000 LinkedIn Ads campaign generates, on last-click attribution, 3 closed deals worth €42,000: 2.33x ROAS. Applying a time-decay multi-touch attribution model, the same campaign turns out to have contributed, as the first touchpoint, to 6 more deals closed in the following 5 months through other channels, for a total attributed revenue of €187,000: real ROAS of 10.4x, four times higher than the last-click reading.