The question "Meta or Google" is framed wrong. They do different jobs, and most DTC brands need both. The real question is how to split budget and what to expect from each.
What Meta does
Meta is a demand-creation engine. Nobody opens Instagram looking to buy your product. Meta puts it in front of them anyway, using creative to interrupt and interest a cold audience. This is where volume and creative quality matter most, because you are earning attention that was not already there. It is also where formats like street interview ads earn their keep.
What Google does
Google is a demand-capture engine. When someone searches for your product, your category, or a problem you solve, they have already decided they want a solution. Search and Shopping ads intercept that intent at the moment of highest purchase readiness. The work here is technical: feed optimisation, bidding strategy, query segmentation, and making sure every click has a clean path to purchase.
Why you usually need both
If you run only Google, you are limited to existing demand, and you will plateau when you have captured the people already searching. If you run only Meta, you create demand but leak conversions to competitors who show up when your newly-interested customer goes to Google to check you out. Running both closes that leak: Meta creates the interest, Google catches it on the way to purchase.
On seasonal accounts this is dramatic. For one home and garden brand we scaled into peak season, Google became a strong second channel at 8.5x+ ROAS while Meta drove the discovery that fed it. Neither channel alone would have produced the +66% revenue growth we saw.
How to split the budget
There is no universal ratio, but a useful starting frame:
- Impulse or discovery products (people do not search for them): weight toward Meta, because you must create the demand first.
- Solution or replacement products (people actively search): weight toward Google, because the intent already exists.
- Most DTC brands land near a 60/40 split and adjust based on where incremental spend still returns.
The right answer is the one the data gives you after testing, which is why we measure incrementality rather than trusting platform-reported numbers on either side.