How many ad creatives should you test per week on Meta?
Direct answer
5, 10, 15, 30, or 50 a week, depending on the account: those are the volumes actually delivered at $300, $600, $900, $1,800, and $3,000/month. But the count is downstream of selection, not an input you pick. Testing more without a real kill filter first just produces more noise at a higher price.
Short version
- Delivery runs in five slabs: 5, 10, 15, 30, or 50 creatives/week, at $300, $600, $900, $1,800, or $3,000/month for Creative Strategy (base rate: flat $60/creative/month, or $20/creative/week).
- Cadence changes shape by slab: a single weekly batch at 5 and 10, two drops a week at 15 and 30, rolling drops at 50.
- The Gauntlet (the pre-launch kill filter) kills roughly 80% of scripts before they ever spend a dollar, so what ships weekly is already the survivors, not the raw output.
- A creative test needs four to six weeks of delivery before the read means anything; launching a high weekly volume into an account that can't read results that fast produces noise, not data.
- The UK/AU lending case ran 13,000+ ads analyzed and systematized through a weekly testing and QA pipeline: most creatives die in testing, and the winner is what gets scaled.
More
Why won't anyone just give you a number?
Because the honest number depends on the account, and a framework is safer to publish than a commitment. It's also less useful.
Here's a number. 5, 10, 15, 30, or 50 creatives a week. Those are the actual delivery slabs: $300, $600, $900, $1,800, or $3,000 a month for Creative Strategy, built off a flat base rate of $60 per creative per month ($20/week). Add Data Analysis and each slab costs 30% more: $390, $780, $1,170, $2,340, $3,900. Read that as what gets delivered at each price point, not as a rule for what every brand should run.
Is the weekly count actually the right lever to pull?
No. The industry reflex is "test more creatives": if performance stalls, launch more. That reflex is usually aimed at the wrong problem.
The weekly count is an output of how much survives selection, not an input you choose. I kill about 80% of scripts before they ever spend a dollar. That's the Gauntlet, the pre-launch filter that separates what gets tested from what gets written and binned. An account testing 30 creatives a week isn't necessarily better resourced than one testing 10 — it might just have weaker selection upstream, generating volume to compensate for a filter that isn't doing its job. Pushing more volume through a weak filter doesn't fix it; it pays extra to confirm the filter was already weak.
Why does the cadence change shape at 15 and again at 50?
Because the delivery mechanics change, not just the count. At 5 and 10 creatives a week, one batch a week is enough; there's time to build, review, and ship in a single cycle. At 15 and 30, delivery splits into two drops a week, because holding 15+ creatives for one Friday batch means some of them sit finished for days doing nothing. At 50, it moves to rolling drops: creative ships as it's ready rather than on a fixed schedule, because a weekly cliff at that volume creates its own bottleneck.
The number that should worry you more than the weekly count is the read time. A creative test needs four to six weeks of delivery before the result means anything. If you're launching 30 a week into an account that needs four to six weeks per read, you're generating volume faster than you can interpret it — noise, dressed up as a testing program. The UK/AU lending account runs a weekly testing and QA pipeline against 13,000+ ads analyzed and systematized, and the operating description is blunt: most creatives die in testing, and the winner is what compounds. That's true whether the weekly count is 5 or 50. Volume doesn't change the ratio. It just changes how fast you burn through it.
Proof
| Creatives/week | Creative Strategy /mo | Creative Strategy /wk | + Data Analysis /mo | + Data Analysis /wk |
|---|---|---|---|---|
| 5 | $300 | $100 | $390 | $130 |
| 10 | $600 | $200 | $780 | $260 |
| 15 | $900 | $300 | $1,170 | $390 |
| 30 | $1,800 | $600 | $2,340 | $780 |
| 50 | $3,000 | $1,000 | $3,900 | $1,300 |
Source: the published pricing on sachitbansal.com. Base rate: a flat $60/creative/month ($20/creative/week); "+ Data Analysis" adds a flat 30% on top.
This is what gets delivered at each price point, not an industry benchmark for how many creatives any given brand should run. What a specific account should test depends on how much survives its own selection process and how fast it can read a result, not on matching someone else's slab.
FAQ
Is there an industry-standard number of creatives to test per week?
Not one I'll state. I have no figure I can stand behind for "most brands," and every published framework that claims one is guessing. The slabs above are what gets delivered at five real price points. That's a fact about my delivery, not a benchmark for yours.
What happens if I test more than I can read in time?
You get volume without insight. First ads are live in week two, but a creative test needs four to six weeks before the read means anything. Before that you're looking at noise and calling it a trend. Running 50 a week into an account you can only read every six weeks means most of that volume never gets a fair verdict.
Does a higher weekly count mean a better testing program?
Not by itself. The count reflects delivery capacity and account needs, not testing quality. A smaller weekly batch that survives a strict pre-launch filter can outperform a larger one that doesn't, because the smaller batch already spent its losses before touching the ad account.
Next step
If you want a real read on the right weekly volume for your account, that's exactly what the diagnostic checks. Book the 20-minute diagnostic
Book the 20-minute diagnostic