How many ad creatives should you test per week? The numbers, by spend tier.
By Denzel Geng · Updated October 25, 2026
Facts on this page verified October 25, 2026 against public sources.
Most answers to this question are a round number somebody likes. This page uses the two datasets that disclose their sample — Motion's 578,750 creatives across 6,015 Meta accounts, and Common Thread Collective's 504 stores and $3.35 billion of spend — plus Meta's own limits on its testing tool, and shows how to turn them into a number for your account.
On Meta, accounts spending under $10K a month launch a median of 2.8 new creatives a week; $10–50K accounts launch 4.1; $50–200K launch 6.6; $200K–1M launch 11.2; and $1M+ accounts launch 18.8 (Motion, 578,750 creatives across 6,015 accounts, September 2025 to January 2026). The top quartile in each tier ships 1.7 to 2.9 times that: 4.8, 8.0, 15.9, 31.1 and 54.6 a week. Roughly 5% of creatives become winners, so the number you should test is set by how many winners you need each month, not by a rule of thumb.
See the concepts winning in your niche- How many creatives per week do Meta advertisers actually launch?
- What counts as a winning creative, and how many win?
- Does the number change by industry?
- Which formats win most often?
- How to work out how many creatives you should test
- How many ads can you test at once in Meta's creative testing tool?
- Why do other guides give such different numbers?
- The bottleneck is concepts, not production.
How many creatives per week do Meta advertisers actually launch?
The only dataset that answers this with a disclosed sample is Motion's Creative Benchmarks 2026: $1.29 billion of realised Meta spend, 578,750 unique creatives and 6,015 advertiser accounts (Facebook and Instagram), September 1, 2025 to January 1, 2026. The figures are medians across the accounts in each monthly-spend tier, with the top quartile of each tier beside them.
| Monthly Meta spend | Median new creatives / week | Top 25% of accounts | Winners / month (median → top 25%) |
|---|---|---|---|
| Micro · under $10K | 2.8 | 4.8 (1.7×) | 0.0 → 0.0 |
| Small · $10K–50K | 4.1 | 8.0 (2.0×) | 0.2 → 0.5 |
| Medium · $50K–200K | 6.6 | 15.9 (2.4×) | 0.7 → 2.0 |
| Large · $200K–1M | 11.2 | 31.1 (2.8×) | 1.7 → 5.9 |
| Enterprise · $1M+ | 18.8 | 54.6 (2.9×) | 3.9 → 10.4 |
Source: Motion, Creative Benchmarks 2026 — "Top 25% of accounts" (last updated April 17, 2026; read October 25, 2026). Means run higher than medians because a few accounts ship a lot: 3.3, 5.1, 9.3, 16.6 and 29.2 a week for the same tiers. Tiers are monthly Meta spend; accounts with fewer than 10 unique creatives were excluded (methodology).
What counts as a winning creative, and how many win?
Motion's winner is a creative whose spend reaches at least 10 times the account's median creative spend and at least $500 — a spend-allocation definition, not ROAS. By that measure about 5% of creatives win, and the rate rises with scale: about 3.8% in Micro accounts up to about 8.2% in Enterprise accounts. Half of all creatives (50–53%) are losers that get turned off, or never reach spend, before day 28. Roughly 6% of ads in any given account end up carrying the majority of that account's spend.
The point Motion makes about its own table is the one that matters for the volume question: top-quartile accounts don't have dramatically higher hit rates. They have more winners because they ship more. In Motion's words, they are "producing dramatically higher absolute winner counts because they're shipping more." A Medium account launching the median 6.6 creatives a week gets about 0.7 winners a month; the top quartile launching 15.9 gets about 2.
Common Thread Collective's cut of 504 stores ($3.35 billion of Meta spend over 365 days, 7-day-click attribution, published July 30, 2026) describes the same funnel from the spend side: 74% of ads received any spend, 24% cleared $100, 7.2% cleared $1,000, 1.8% cleared $5,000, and 0.9% became what CTC calls whales — "roughly one in a hundred." CTC's top advertisers ship 12–19+ new evergreen creatives a week, with hit rates "more than double those of smaller accounts." Catalog ads are excluded from that math; they carry spend almost by default.
Does the number change by industry?
Yes, mostly at the top. Motion's vertical cut shows the same tiers for the three verticals most DTC teams sit in. Within a tier the medians are close; the spread opens up at $200K+ a month.
| Vertical (median creatives / week) | Micro | Small | Medium | Large | Enterprise |
|---|---|---|---|---|---|
| Health & Wellness | 3 | 4 | 11 | 19 | 46 |
| Fashion & Apparel | 3 | 5 | 12 | 18 | 33 |
| Beauty & Personal Care | 3 | 4 | 8 | 15 | 26 |
Source: Motion, "Testing by vertical". Values are median creatives launched per week per account within each vertical × tier cell; verticals with fewer than 50 accounts were pooled into "Other."
Which formats win most often?
The cheapest formats have the highest hit rate, which is part of why volume is possible at all. Motion's on-stage breakdown of the same dataset, by creative format:
| Format | Hit rate |
|---|---|
| Text only | 11.6% |
| Product image + text | 8.75% |
| UGC | 7.56% |
| High production | 6.87% |
Source: Motion, "Meta ads in 2026: how many creatives do you actually need to launch?" (recorded March 2026). A higher hit rate for text-only ads does not mean text-only ads make more money; it means a cheap format earns a test slot more often than it is given one.
How to work out how many creatives you should test
Start from the winners you need, not from the benchmark. The benchmark tells you what the median account in your tier does; the median account in every tier gets fewer than two winners a month.
1. Decide how many winners you need per month
A winner is a creative that can carry real spend for weeks. Most accounts need one or two a month to keep the account from running on a single fatiguing ad. If you plan to scale budget, you need more, because every new budget level needs creatives that work at that level.
2. Use your tier's hit rate
Under $10K a month, plan on about 4% of creatives winning; $10–50K about 6%; $50K–200K about 8%; $200K and up 8–9%. Your own account history is better than any of these once you have 50+ creatives to count.
3. Divide
Creatives per month = winners needed ÷ hit rate. Two winners a month at a 6% hit rate is about 33 creatives a month, or 8 a week — which is exactly where Motion's top-quartile Small accounts sit (8.0 a week). One winner a month at 4% is 25 a month, about 6 a week, double the Micro median.
4. Count concepts, not files
Twenty crops of one idea are one creative for this purpose. Meta's own advice is to diversify by concept, and distinct concepts reach people a near-copy can't. If your 8 a week are two angles with four hooks each, you are testing two things. The creative testing playbook covers the concept-versus-iteration split and the test structure.
5. Check the number against your production capacity
The top-quartile numbers above are only possible because the cheap formats (text-only, product image, UGC) carry most of the volume. If every creative is a produced video, your realistic number is the median for your tier, and the lever is a higher hit rate from better angles rather than more files.
How many ads can you test at once in Meta's creative testing tool?
Between two and seven. Meta's creative testing tool, set up inside an existing campaign, compares 2 to 7 test ads built as copies of an ad, not existing ads; Meta suggests using no more than 20% of the campaign's existing budget on the test, and it works only with the Highest volume bid strategy (help page read October 13, 2026). That is a per-test limit, not a weekly volume: a Medium account at the median 6.6 creatives a week is running roughly one such test a week, and a top-quartile one is running two or three in parallel or rotating concepts through ABO ad sets instead. The structures that scale past the tool's limit are in the creative testing playbook.
Why do other guides give such different numbers?
Because most of them have no dataset. The ranges that circulate — 5 to 8 creatives a month at $5K, 20 to 30 at $30K, 60+ at $150K, or a flat "15 to 25 new creatives monthly" — come from agency planning guides that say so themselves ("planning ranges, not guarantees") and cite no sample. They are not wrong as a way to start; they are just someone's habit, and the monthly versions imply far less than the weekly medians in Motion's data at the same spend. When a number has no sample size next to it, treat it as a rule of thumb.
Two other things to keep separate from the volume question. Hit rate is not the goal — Motion's spend-based winner definition says nothing about ROAS, and both Motion and CTC say so. And refresh cadence is not the same as test cadence: Motion's loser definition (turned off before day 28) tells you how fast losers get cut, not how long winners last. Fatigue timing is covered under ad fatigue.
Common questions.
It depends on spend. Motion's 2026 benchmark of 578,750 creatives across 6,015 Meta accounts puts the median at 2.8 new creatives a week under $10K a month, 4.1 at $10–50K, 6.6 at $50–200K, 11.2 at $200K–1M and 18.8 at $1M+. The top quartile of each tier ships 4.8, 8.0, 15.9, 31.1 and 54.6. Work backwards from the winners you need: at a 5–8% hit rate, two winners a month needs roughly 25–40 creatives a month.
Multiply the weekly median by about 4.3: roughly 12 a month under $10K in spend, 18 at $10–50K, 28 at $50–200K, 48 at $200K–1M and 81 at $1M+. Top-quartile accounts run about double. Common Thread Collective's data on 504 stores puts top advertisers at 12–19+ new evergreen creatives a week, which is 50–80 a month.
About 5%. In Motion's dataset, where a winner is a creative reaching 10× the account's median creative spend and at least $500, the hit rate runs from about 3.8% in accounts under $10K a month to about 8.2% in accounts over $1M. Common Thread Collective's spend funnel across 504 stores found 7.2% of ads cleared $1,000 in spend and 0.9% became outsized "whales." About half of all creatives are turned off before day 28.
Both, but the data says volume is what separates the top quartile. Top-25% accounts in Motion's benchmark don't have much higher hit rates than the median account in their tier; they ship 1.7–2.9× more creatives and therefore get 2–3× more winners. The practical version: keep the hit rate up with distinct concepts, then ship each concept in several cheap executions.
Meta's creative testing tool allows 2 to 7 test ads, built as copies, using no more than 20% of the campaign's budget and only with the Highest volume bid strategy. Outside the tool, the common structure is one concept per ABO ad set with 3–6 variations, cutting losers after a few days; the creative testing playbook covers the structures and budgets.
The median account under $10K a month launches 2.8 new creatives a week, and the top quartile launches 4.8 — so 3 to 5 a week, built as two or three distinct concepts rather than variations of one. At a roughly 4% hit rate that yields about one winner every two months, which is why small accounts should spend their effort on the angle (the cheapest way to raise hit rate) and on cheap formats like statics and text-led ads.
The benchmarks on this page measure launch rate, not lifespan. What Motion's data does say is that half of all creatives are turned off before day 28, so a creative that survives a month is already in the better half. Winners run until their own metrics decay; see the ad fatigue glossary entry for the signals to watch.
Pick the winners you need. The volume follows.
The accounts shipping 15 a week aren't inventing 15 ideas.
They run a few proven angles through many cheap executions. ViralMojo finds the angles and hooks winning in your niche this week — organic and paid — decodes why they work, and writes the brief and scripts, so your weekly number is a production schedule instead of a blank page.
- This week's winning ads and organic posts in your category, decoded to hook, angle and structure
- A brief per angle that names the test variable
- Hooks and scripts written in your customers' own words, ready to shoot
- Works from your product URL — no ad account required
Plans from $99/mo, flat — see ViralMojo pricing.