Split a client budget by the job each platform does. Google captures demand that already exists, and Meta creates demand that has not shown up yet. Start with the size of the search demand, fund Google up to where extra spend stops finding new searches at a fair cost, put the rest into Meta, and re-check the split every week against one combined number.
That is the whole method. The rest of this post is how to run it without guessing. If you manage a client with a daily budget that has to land close to plan, the split between Google and Meta is not a one-time decision. It is a routine.
Start with the job each platform does
Most arguments about Google versus Meta go nowhere because they treat the two as rivals for the same click. They are not.
Google captures demand
Someone types a query. They already want something. Your ad shows up at the moment they are looking. The ceiling here is the number of people searching. You cannot buy more searches than exist. When a search campaign hits that ceiling, extra budget buys worse queries, higher costs, or nothing at all.
Meta creates demand
Meta shows your client to people who were not looking. Some of them will want it, and some of them will search for it later. The ceiling here is much higher, but the path from impression to sale is longer and noisier. Meta tends to need good creative and enough volume to learn.
Why this changes the split
If the client sells something people already search for in large numbers, Google earns a bigger share early. If the client sells something new, something people do not know to search for, or something with a small search pool, Meta has to do more of the work, and Google's share stays small because it cannot absorb more.
Here is a quick test. Look at how much of the available search demand your Google campaigns already capture. If you are showing on most of the searches that matter and the budget is still not spent well, the problem is not Google's share. The ceiling is the issue, and the next dollar belongs somewhere else.
Set the opening split
You need a starting ratio. It will be wrong. That is fine, because the point is to have a ratio you can test.
Step 1: Size the search pool
Check how much search volume the client's offer has and how much of it your current Google campaigns reach. If the pool is large and you are only reaching part of it, there is room to add budget. If you already reach most of it, Google is near its ceiling.
Step 2: Set a target for each platform
Use the client's real target. If the client works to a target CPA, set it per campaign where it matters, with an account-level target as the fallback. If the client works to ROAS, do the same. Judge Google and Meta against the same target so the comparison is fair.
Step 3: Fund the proven channel first, then test the other
Put enough on the platform with the clearest path to the target to hit it. Put a smaller, fixed amount on the other so it can learn. Do not starve either one to the point where the data means nothing.
A worked example
This is an example with round numbers, not a benchmark. Say a client has a monthly budget of $30,000 and a $50 target CPA on both platforms.
- Google search has a modest pool. Your campaigns reach most of it, and they run at about the target.
- Meta has not been run at scale for this client, so you do not know its real cost per result yet.
A sensible opening split is $18,000 on Google and $12,000 on Meta, which is 60 and 40. After two weeks you look at the result. Say Meta is running under the target and has room to take more, while Google's extra spend is coming in above the target because the cheap queries are used up. You move a slice, say $3,000 a month, from Google to Meta. Then you wait and look again.
You did not decide Meta was better. You let each platform show where the next dollar works hardest.
What each platform's own recommendations are built to do
Both platforms will tell you what to do with your budget. Read those suggestions as input, not as instructions.
Google's recommendations tend to point toward more spend inside Google, such as broader match types or higher budgets. Meta's tend to point toward more spend inside Meta, such as wider targeting or a bigger daily budget. That is not a trick. Each platform is built to improve results inside its own walls, and neither can see the other's numbers.
So when a platform suggests raising a budget, ask a second question: what does the combined result say? If Google says to raise the budget and the combined cost per result gets worse when you do, the suggestion was right for Google and wrong for the client.
This is also why a single source of truth matters. When you judge both platforms as one budget, you remove the pull toward whichever dashboard you happened to open last.
Look at the combined number, not the platform number
Platform-reported conversions overlap. A buyer who saw a Meta ad and later searched on Google can show up in both reports. Add them together and the total will often be higher than the client's real sales or leads.
The fix is to keep a combined view. Compare spend across all platforms with what the business actually earned. The marketing efficiency ratio, or MER, does this by dividing total revenue by total ad spend. Use platform numbers to see what is happening inside a platform, and use the combined number to decide whether the split is working.
Re-split every week
A split that is right in week one drifts. Seasonality, new creative, competitors, and the client's own offers all change what each platform can do.
A weekly routine you can run in under an hour
- Check the combined result first. Is total spend landing close to plan, and is the combined cost per result or MER where it should be?
- Check each platform against its target. Which one is under, which is over, and by how much?
- Check for a ceiling. Is Google spending its budget on weaker queries? Is Meta's frequency climbing?
- Check what changed. New creative, a new offer, a budget edit, a tracking change. If something changed, do not judge the platform yet.
- Move money in small steps. Shift a modest slice toward the platform with room and away from the one at its ceiling. Never move so much that you cannot tell what caused the result.
- Write down what you moved and why. Next week you will want to know.
Signals that say move money
- One platform is well under target and has not hit a spending ceiling, while the other is over target.
- A Google campaign is limited by budget while its cost per result is at or below target.
- Meta frequency is rising and cost per result is getting worse on a creative that used to work. This points to fatigue, so the first fix is fresh creative, and only then a budget move.
- Combined MER is falling even though the platform reports look fine, which usually means the platforms are claiming the same sales.
Signals that say wait
- A campaign was edited in the last few days.
- A platform is still in a learning phase.
- One bad day with no pattern behind it.
- A conversion tracking issue you have not fixed yet. Fix the tracking first, because no split can be right on bad data.
Watch the small accounts, too
Multi-location clients make this harder. When a client has dozens of Google and Meta accounts, each on its own daily budget that has to land close to plan, one expired card on one account can go unnoticed for two days. The split between platforms can look fine on paper while one account is not spending at all. Part of the weekly routine is simply checking that every account is delivering what it should.
What about TikTok?
TikTok can earn a place in the split, but it should earn it the same way the others do. Give it a small, fixed test slice, set a target that fits how you will judge it (link cost per click is common for a first read), and compare it on the same combined view. If it hits the target and has room, it gets more. If not, it goes back to zero without any damage to the rest.
Where Ads Arsenal fits
Ads Arsenal does not sell media, so it has no stake in which platform wins. It reads Google, Meta, and TikTok as one budget, and its Overview shows combined metrics across platforms with MER at the top. You set an account-level target as the fallback, set a CPA or ROAS target per campaign where it matters, and write rules for groups of campaigns, for example judging every TikTok campaign on link cost per click. Platform recommendations are read as inputs, not followed blindly.
It reads your accounts all day and ranks what to change in an Insights queue, with the reasoning and the numbers behind each recommendation in plain English. Nothing changes in an account until someone approves it. Approve a change and it is written to the account, checked a few minutes later to confirm it landed, and tracked to see whether it worked. Any change rolls back with one click. That makes the weekly routine above faster to run, and it leaves the decisions with you.
FAQ
Is Google Ads or Facebook Ads better for my client?
Neither is better in general. Google is better at capturing people who are already searching for what your client sells. Meta is better at reaching people who have not started looking yet. The right answer depends on how much search demand exists for the client's offer and what each platform costs per result against the client's target.
How should I split a budget between Google and Meta?
Start with the size of the existing search demand. Fund Google up to the point where more spend stops finding new searches at an acceptable cost, then put the rest into Meta. Treat the first ratio as a guess, and let two weeks of results against the client's target move it.
How often should you move budget between platforms?
Review weekly and move in small steps. Daily moves react to noise, and monthly moves leave money in the wrong place for too long. If a platform is learning or a campaign just changed, wait for it to settle before judging it.
Why do Google and Meta both claim the same conversion?
Each platform counts conversions using its own attribution rules and its own view of the customer. The same buyer can see a Meta ad, search on Google later, and be claimed by both. Add the platform numbers together and you will usually count more conversions than the client actually got.
Should TikTok get part of the budget?
Only if the client has a reason and the budget to learn from it. Give it a small, fixed test slice with its own target, judge it on the same combined view as the other two, and keep it only if it earns its share.
Try it on your own accounts
Ads Arsenal is in a free design-partner beta for agencies and in-house teams running real spend on Google, Meta and TikTok, with Search Arsenal on the same login. While it is in closed beta, Meta connects when you sign in with your Meta account, and Google connects through a manager-access request we set up with you on a short call. The client keeps owning the ad account. Become a design partner.