Google vs Meta Budget Split Calculator

Split one monthly budget across Google and Meta — and find out whether you should split it at all. The tool works from Meta’s published learning-phase threshold, so it will tell you plainly when the honest answer is Google only.

This is the interactive companion to Google or Meta ads first, by budget.

Inputs

Total monthly ad budget

Media spend only, across both platforms. Illustrative default: $2,000.

Target cost per acquisition

What one lead or purchase is allowed to cost. Illustrative default: $60.

Optimizing for

The arithmetic is identical either way — this only changes which event Meta needs 50 of.

Results

Do not split — run Google only until about $17,333/mo

100% Google

$2,000/mo

At a $60.00 target CPA, one Meta ad set needs about $13,000/mo to average ~50 leads per 7 days and exit the learning phase. You are $11,000/mo short of that, so a 50/50 split would leave Meta stuck in learning and Google underfunded at the same time. Search demand already exists, so spend the whole budget where the intent is, and revisit Meta at $17,333/mo.

Meta learning floor

$13,000

Monthly spend one Meta ad set needs to average ~50 leads per 7 days.

Split threshold

$17,333

Meta's floor ÷ 0.75 — below this, funding Meta leaves Google under a quarter of the budget.

Meta weekly floor

$3,000

Target CPA × 50 leads in a 7-day window.

Google

100%

$2,000/mo

$65.75/day

The arithmetic behind the verdict

Meta’s published learning-phase guidance is roughly 50 optimization events per ad set per 7 days — Meta Business Help Center. General guidance, not a guarantee.
Weekly floor = $60.00 target CPA × 50 leads = $3,000.00 per week.
Monthly floor = $3,000.00 × 52 ÷ 12 = $13,000.00 per month.
Split threshold = $13,000.00 ÷ 0.75 = $17,333.33 per month — the point where funding Meta still leaves Google 25% of the total.
Your budget of $2,000.00 is below that threshold, so the recommendation is put 100% on Google for now.
Daily budgets are monthly ÷ 30.4 (that is 365 ÷ 12, the average month length — plain arithmetic, not a platform figure).
At this CPA your total budget could cover the learning floor for 0 Meta ad sets — the 50-event guidance applies per ad set, not per account.

The 25% Google-share rule is this tool’s own convention, not a platform figure. Only the ~50 events per 7 days comes from Meta. We publish no customer performance benchmarks.

The split is decided by Meta’s learning phase

Not by a percentage rule of thumb. One platform publishes a minimum event volume; the other does not.

Meta publishes a ~50-event threshold

Meta's Business Help Center says an ad set generally leaves the learning phase after roughly 50 optimization events in a 7-day window. That published figure, not a preference, sets the minimum Meta budget this tool works from.

Learning floor = CPA × 50 × 52 ÷ 12

Fifty events a week at your target CPA is a weekly spend figure. Multiply by 52 weeks and divide by 12 months and you get the monthly spend one Meta ad set needs before its results are worth reading.

Below the floor, Google is the honest answer

Google search publishes no equivalent event threshold gating delivery, and search demand already exists. If you cannot fund Meta's floor, a 50/50 split underfunds both platforms instead of one working one.

Daily budgets are what you actually type

Both platforms take a daily number. This tool divides each monthly allocation by 30.4 days — that is 365 ÷ 12, plain arithmetic, not a platform figure.

Where the 50-event number comes from

Source: Meta Business Help Center. Meta documents the learning phase and states that an ad set generally leaves it after roughly 50 optimization events in a 7-day window. An optimization event is whatever you told the ad set to optimize for — a lead, a purchase, a signup. Meta presents this as general guidance about how delivery stabilises, not as a guarantee, and it is the only external figure this calculator uses.

Turn that into money and the budget question answers itself. Fifty events a week at your target CPA is a weekly spend requirement, and multiplying by 52 ÷ 12 turns it into the monthly number you actually plan around. At a $40 target CPA that is about $8,667 a month for a single ad set; at $120 it is $26,000. Most small budgets are nowhere near it, and that is the useful finding, not a failure of the tool.

Google search does not work the same way. Someone typing your service into the search bar already has intent, and Google publishes no comparable event threshold gating delivery, so a smaller budget can still produce a coherent test. That asymmetry — not any claim that one platform performs better — is why the honest recommendation below a certain budget is Google first.

If you only want the Meta side of this question — how many ad sets a given budget can actually support — the Meta learning phase calculator uses the same formula without the Google allocation on top.

Once both platforms are funded and running, stop using this formula and start using measured results. Take real spend and conversions into the CPA calculator and move budget toward whichever platform is buying outcomes cheaper.

How AdFlint handles the split after launch

AdFlint runs Google and Meta from one place, tracks spend and conversions on both, and shifts budget toward whichever platform is producing outcomes at a lower cost — so the starting split on this page stops mattering once real data exists.

AdFlint’s self-serve plans are a flat monthly fee with 0% markup on ad spend — your card pays Google and Meta directly. They start at $10/mo for up to $100/mo of ad spend, then $40 up to $1,000, $89 up to $2,500, $109 up to $5,000, $119 up to $10,000, $219 up to $20,000, and $319 up to $30,000, with custom pricing above that. Every plan starts with a 7-day free trial. The separate done-for-you Managed tier is priced differently: a $249/mo floor plus a share of managed spend.

Try AdFlint free AdFlint for small business See pricing

Questions

How should I split my ad budget between Google and Meta?

Start from Meta's learning-phase floor rather than from a percentage. Meta's Business Help Center says an ad set generally needs roughly 50 optimization events in a 7-day window to exit the learning phase, so the monthly spend one ad set needs is approximately target CPA × 50 × 52 ÷ 12. If funding that floor would leave Google less than a quarter of your total budget, put everything on Google. If it leaves Google more than that, fund Meta's floor first and put the remainder on Google, then rebalance on measured CPA once real data exists.

Why does the calculator sometimes tell me not to split at all?

Because splitting a small budget produces two underfunded platforms instead of one working one. At a $60 target CPA, one Meta ad set needs roughly $13,000 a month to average 50 events a week. A $2,000 budget split 50/50 leaves Meta permanently in the learning phase and Google at half strength. The honest answer at that budget is Google only.

Where does the 25% figure in the split threshold come from?

From us, not from any platform. It is this tool's own convention: we only recommend splitting if, after funding Meta's floor, Google still keeps at least a quarter of the total budget. That makes the split threshold Meta's monthly floor ÷ 0.75. Pick a different share and the threshold moves — the arithmetic is shown on screen so you can check it against your own judgement.

What is Meta's learning phase?

It is the period after you create or significantly edit an ad set, while Meta's delivery system is still working out who to show the ad to. Meta documents the exit condition as roughly 50 optimization events in a 7-day window per ad set. Performance during the learning phase is less stable, and editing the ad set restarts it. Meta describes this as general guidance, not a guarantee.

Does the 50-event rule apply per campaign or per ad set?

Per ad set, according to Meta's documentation. That matters for budgeting: three ad sets need roughly three times the floor, which is why splitting one small budget across several audiences usually leaves all of them in learning.

Does the answer change if I optimize for purchases instead of leads?

The arithmetic is identical — 50 events is 50 events. What changes in practice is the cost of each event, since a purchase usually costs more than a lead, which raises the floor. If your purchase volume cannot realistically reach the threshold, a common workaround is to optimize for a higher-funnel event, which lowers the CPA you would enter here but also changes what you are buying.

Should I trust this as a benchmark?

Only the ~50 events per 7 days figure comes from Meta, and Meta frames it as general guidance rather than a guarantee. The 25% Google-share rule is our own convention. Everything else is arithmetic on numbers you entered. We publish no customer performance benchmarks, because we have none to publish.