When Meta and Google are both showing positive ROAS, the instinct is to scale both proportionally. The problem: they are not operating independently. They are measuring the same conversions through different attribution lenses and both claiming credit for the same revenue. Understanding what each platform actually does in the funnel is the prerequisite to a rational budget allocation.
How each platform sees the funnel differently
Meta is a demand creation platform. It shows ads to people who are not looking for your product, it is manufacturing intent by interrupting people in a feed. The conversion path is longer because the user was not in the market when you found them. Google Search is a demand capture platform. It shows ads to people who are actively searching, the intent already exists and you are inserting yourself into a decision that is already in progress. This is why Google typically shows higher ROAS: it is not a better platform, it is reaching a different, and more ready, audience. The user who searches your brand name on Google is often someone Meta warmed up three days earlier.
The attribution overlap problem
A user sees your Meta ad on Monday. On Thursday, they Google your brand name and click your branded Search ad. They convert on Thursday. Google last-click attribution says: Branded Search, 100% credit. Meta last-click attribution says: Meta, 100% credit (within 7-day click window). You have double-counted the same conversion in two different reporting dashboards. Your blended ROAS across both platforms is therefore overstated. The question you need to answer is not "which platform has higher ROAS?" It is "which platform would I need to cut before revenue dropped, and by how much?"
The incrementality test for each channel
Run a holdout test for Meta: remove a control group from all Meta exposure (not just retargeting, all Meta) and measure whether their conversion rate drops. The delta is Meta's true incremental contribution. Then run the same test for branded Search: remove branded Search ads for a portion of users and measure whether organic search captures the traffic. Most brands find that 60–80% of branded search clicks would have happened via organic if the paid ad were absent. Branded Search has a very low incremental ROAS at most companies, you are paying for conversions that would have happened anyway via the organic listing directly below the paid ad.
The allocation framework that follows
Meta budget: sized by the incremental pipeline it generates per pound of spend, not last-click reported ROAS. This number typically justifies more Meta than most companies are running. Google non-brand: sized by actual search volume, demand capture is roughly linear, so if volume is there, you should capture it. Google brand: sized to maintain share of voice, not optimised for ROAS, it is capturing organic intent, not creating new demand, and the ROAS figure is meaningless as a signal. The reallocation that typically follows this analysis: cut branded Search by 40–60%, redirect that budget into Meta prospecting, and measure incremental revenue over 90 days.