The short version
- Your weekly number is the lower of two limits: how many ads your spend can judge, and how many good ads your team can actually make.
- Measure the making limit from the time your last ten ads really took, then take off a fifth for rejects.
- Count ideas, not files. One design in three sizes is one ad.
- Split each week into explore, follow-up, reshape and wild-card slots, and let results move the split.
- Change the pace two or three ads at a time and keep a five-number log, so each week teaches the next.
The short answer
Make as many new ads each week as the lower of two limits allows: how many ads your spend can give a verdict to, and how many good ads your team can finish without cutting corners. For a small brand that usually lands between 3 and 10 a week, and for a large account with a creative team, 20 or more. Pick a pace you could hold for the next eight weeks, because a steady flow teaches you more than a burst of 60 ads followed by a quiet month.
Those ranges are our planning starting points, not industry benchmarks. The rest of this guide is about the second limit, the one most plans skip, and how to measure it for your own team.
Two limits set your weekly number
Think of the weekly number as the place where two limits meet.
The judging limit. A new ad needs enough spend to show whether it sells. Launch more than your testing money can judge and most ads get a handful of impressions each, so nothing earns a verdict. The sum behind this limit lives in our guide on how many ad creatives to test, and the creative test calculator works it out from your own target cost per purchase. This page takes that figure as given and looks at the other side.
The making limit. This is the limit most plans skip. It is the number of finished, checked, on-brand ads your people and tools can turn out in an ordinary week, with orders, support and everything else still going on. A founder who designs ads late on a Sunday can hit 15 once. Hitting 15 every week is a different job.
Your weekly number is whichever limit is lower:
- Can judge 10, can make 4. The fix is production: better inputs, layouts you reuse, a fixed making day or a tool. More testing money would just sit idle.
- Can make 20, can judge 6. The extra ads aren't wasted, but they belong on the bench for the following weeks, not in this week's launch.
Measure your making limit
Most teams guess this number, and most guess high. A short exercise gives you the real figure:
- Time your last ten ads. Count every step, not only the design: finding the photo, writing the line, resizing, getting sign-off, fixing the typo someone spotted after export. Divide the total by ten.
- Count the hours you really have. Take an ordinary week and subtract meetings, packing, customer emails and anything else that can't move. Use what's left, not what you wish you had.
- Divide hours by time per ad. That is your making limit today.
- Take off a fifth for rejects. Some ads won't pass your own review and a few may not pass Meta's. Planning for them keeps the week from overrunning.
How the ads get made changes the answer a lot. The table compares the main routes, using Tidewell's own timings as a filled-in example. Replace them with yours.
| How the ad is made | Tidewell's time per finished ad | What slows it down | Best used for |
|---|---|---|---|
| Photo shoot, then design | About a day for a set of five | Booking, props, retouching | Fresh hero images once a quarter |
| Designer working from existing photos | About 90 minutes | Revision rounds, resizing | Statics with a clear, proven layout |
| Creator video in a UGC style | Two to three weeks from brief to file | Shipping product, filming, edits | Demos, reactions and talking to camera |
| AI batch from one packshot | One batch of up to 12, plus review | Checking every ad by eye | Many angles on one product at once |
Creator videos arrive slowly, so order them weeks ahead and treat each one as a planned slot. Statics and AI batches are what fill the weekly number. Our page on AI UGC ads covers the creator-style route in more depth.
Decide what counts as one ad
Before you set a target, agree with the team on the unit. A target counted in exported files is easy to hit and teaches little, and it also hides where the making time really goes.
- Sizes are export work, not ads. Turning a 4:5 design into 1:1 and 9:16 takes Tidewell's designer about 15 minutes. Log it as part of the ad, not as two more.
- A colour swap is not a slot. If a viewer would barely tell two versions apart, they share one place in the plan.
- A new opening line on a proven image is a variant. It is cheap to make and still changes what people take in during the first second, so it earns a slot.
- A new reason to buy is a concept. A price comparison, a moment of use and an answered complaint each need their own brief, and they take the longest to make.
There is a delivery reason for this too. In a post dated December 2, 2024 (opens in a new tab), Meta's engineering team described Andromeda, a retrieval stage that selects from tens of millions of candidate ads down to a few thousand relevant ones, and wrote that it expects the number of ad creatives in its systems to grow significantly. It sets no weekly quota for advertisers. What it means for your counting is covered in our Andromeda guide.
So "eight ads this week" should mean eight briefs' worth of creative, each with its sizes, not three designs exported in three shapes.
Give every slot a job
A weekly number is easier to hit, and more useful, when each ad has a job before anyone starts making it. Split the week into four kinds of slot:
- Explore slots hold new concepts: an angle, a format, a setting or a person you haven't tried.
- Follow-up slots hold variants of a recent winner, such as a new opening line, a new scene or the same idea moved to a different moment of the day.
- Reshape slots move a winner into another format: a short 9:16 video built from a winning 4:5 static, or a carousel that breaks a winning claim into steps.
- One wild card for the idea the team keeps arguing about. It costs a single slot and stops the plan from only ever testing safe ideas.
How you fill the slots depends on what the account has already taught you:
| Where the account is | Explore | Follow-up | Reshape | Wild card |
|---|---|---|---|---|
| No winner yet | Nearly all | None | None | 1 |
| One or two winners | About a third | About half | 1 | 1 |
| Several winners, some slowing down | About half | About a third | 1 or 2 | 1 |
For an eight-ad week with one winner, that reads as three new concepts, three follow-ups, one reshape and one wild card.
The split also changes your making limit, because the slots don't take the same time. Tidewell's 90 minutes per static is an average. A follow-up that keeps the winner's photo and layout and changes one line takes its designer about 30 minutes. A reshape into a short video takes about an hour. A new concept, with a fresh layout and a new scene, takes closer to two and a half hours. Run Tidewell's 10 hours through that:
| Week type | Slots | Making time | Fits in 10 hours? |
|---|---|---|---|
| No winner yet | 3 explore, 1 wild card | 10 hours | Just, with no slack |
| One winner | 1 explore, 1 wild card, 4 follow-up, 1 reshape | 8 hours | Yes, with 2 hours spare |
Seven ads in the second kind of week cost less time than four in the first. That is why a young account should plan fewer ads, not more, and why the number can rise once a winner gives you something to follow up on.
Let results move the split. If follow-ups keep beating new concepts week after week, one idea is carrying the account, and that is the moment to explore harder, before it stops working. Our guide on concept vs iteration testing goes deeper on the balance.
Starting numbers by team shape
Treat these as a first guess, then correct them with your own making limit. They are planning suggestions, not measured averages.
| Team | Weekly number to start with | How the week runs | Weak point |
|---|---|---|---|
| Solo founder | 3 to 6 | One brief, one making session, launch on a fixed day | The week a big order or a sale lands |
| Founder plus a freelancer | 6 to 10 | Founder briefs on Monday, freelancer delivers by Wednesday | Revision rounds eating the freelancer's hours |
| In-house team of two or three | 10 to 20 | Shared idea list, one person owns briefs, one owns files | Sign-off from someone outside the team |
| Agency, per client | 8 to 15 | One brief per client per week, made in batches | Several sale calendars landing at once |
Bigger spenders with a full creative team can go well past 20. At that size the slow part tends to move from making ads to checking, naming and launching them, which is the job of a bulk ad launcher. In BulkAds each brand keeps its own kit, products and competitors, so client work stays separate. See BulkAds for agencies.
A worked production week
Here is Tidewell's week after it added AI batches next to the designer's statics and raised its number to the 8 its testing spend can judge.
| Day | Task | Output | Time |
|---|---|---|---|
| Monday | Read last week's verdicts, write a short brief per slot | 8 briefs | 45 minutes |
| Tuesday | Designer makes 2 statics, one AI batch of 12 from the bottle packshot | 14 drafts | 3 hours of design, plus the batch |
| Wednesday | Review: keep 6 AI ads for this week, bench 3, drop 3 | 8 finished ads | 1 hour |
| Thursday | Upload paused, name, tag and submit for review | 8 ads in review | 30 minutes |
| Friday | Check approvals, fix anything rejected, switch on | 8 live ads | 15 minutes |
A few things to notice:
- The batch covers more than the week needs. Twelve drafts for six slots leaves room to drop ads where the label drifted or two angles came out too alike, and the three good spares start next week's bench. BulkAds' pre-flight check already flags ads where the product doesn't match the original photo, which shortens Wednesday.
- Thursday submission is deliberate. Meta's Advertising Standards (opens in a new tab) in its Transparency Center say review starts automatically before an ad runs, looks at the images, video, text and landing page among other things, and is typically completed within 24 hours, although it can take longer in some cases (as of October 2026). Submitting on Thursday means a slow review still lands before the weekend, not in the following week.
- The money is easy to see. Tidewell pays its designer $9.30 an hour, so the two statics cost $28 of design time. In BulkAds, the batch of 12 image ads uses 24 credits at 2 per image ad, with the total shown before the batch starts. Our guide to what an ad creative costs compares the routes in more detail.
When to raise the number, and when to lower it
Raise it when:
- Every ad reaches a verdict inside a week and the judging limit has room to spare.
- Your best ads are slowing down and the bench is empty. The creative refresh schedule shows how to see that coming.
- A sale season is near. Offer ads for Black Friday or Diwali have a fixed life, so make them in the weeks before, on top of the usual number.
- You are opening a new market or language, which needs concepts written for it, not old ads translated.
Lower it when:
- Ads finish the week with almost no spend. You are making more than you can judge.
- Your own review keeps catching typos, wrong product details or off-brand colours. Speed is costing quality, and Meta's Advertising Standards note that lower-quality ads which don't break any policy may still see weaker performance (as of October 2026).
- Rejections pile up. Read the reason given for each one and fix the brief or the standing rules, not just the single ad.
- Nobody has time to fill in the log. A pace you can't record is a pace you can't learn from.
Move the number two or three ads at a time, not by doubling it. A sudden jump usually shows up as rushed briefs and a pile of near-copies.
Keep the pace without adding people
- Start every week from the log. Last week's winners write half of this week's follow-up briefs.
- Reuse your inputs. One clean packshot can sit behind many scenes and layouts. See one product photo, many ads.
- Start from a layout, not a blank page. The ad templates and the hook bank give each new concept a structure to fill.
- Make in one block. Two hours on one day beats twenty minutes on six days, because the set-up cost is paid once.
- Order slow formats early. If creator videos take weeks, order next month's this week.
- Let the follow-ups draft themselves. In BulkAds, Autopilot drafts four fresh variants of every winner, and when there aren't enough approved ads waiting it drafts a small batch for your approval from patterns that work in your category. Your follow-up slots arrive as drafts, so the half-hour jobs above shrink to a review.
A creative sprint is a good way to build the first bench quickly if you are starting from zero.
A five-number production log
Track five numbers each week in a plain sheet. Here are three weeks of Tidewell's, all fictional:
| Week | Made | Launched | Got a verdict | Winners | Note |
|---|---|---|---|---|---|
| 1 | 10 | 8 | 7 | 1 | "A year of plastic bottles" angle won |
| 2 | 9 | 8 | 8 | 2 | Two follow-ups on that angle beat the original |
| 3 | 12 | 8 | 6 | 0 | Two ads barely spent, the bench now holds 7 |
Read it across the columns:
- Launched often below made means the bench is growing, so you can make less or launch more.
- Verdicts often below launched means you are launching more than you can judge.
- Winners at zero for three weeks means the explore slots need different ideas, not a faster pace.
Questions people ask
How many new ads should a small brand make each week?
For a solo founder, 3 to 6 a week is a sensible start, and 6 to 10 for a founder working with a freelancer. Then check that number against how many ads your spend can judge and how long your ads really take to make.
Is it better to make more ads or better ads?
Better ads, as long as you still learn something every week. A smaller set of distinct, well-made concepts teaches more than a pile of rushed near-copies, and Meta notes that lower-quality ads may see weaker performance even when they break no rules.
Do the square, vertical and full-screen versions of one ad count as separate ads?
Not for planning. They carry one idea to different placements, so count them as one concept and list the sizes separately.
Which day should new ads go live?
Pick one fixed day and work back from it. Meta says ad review is typically completed within 24 hours but can take longer, so submit at least a day before you want the ads running.
How far ahead should ads be made?
Keep about one week of finished ads on the bench. Order creator videos several weeks ahead, and make sale-season offer ads in the weeks before the sale, on top of your usual number.
Can an AI ad tool raise my weekly number?
It raises the making limit, because one batch can cover many angles on one product. It doesn't change how many ads your spend can judge, and every AI ad still needs a human look before launch.
Keep going
- Features
AI ad generator
On-brand image ads in batches of up to 12
- Features
Autopilot
Stops losers, scales winners, inside your caps
- Free tools
Creative test calculator
How many new ads you can test, and how long until every one has a verdict
- Guides
How many ad creatives should you test? The number, and the maths.
A straight answer, then the simple sum that sets your real number, how many new ads to launch each week at your spend, and when to stop testing and scale.
- Guides
How often to refresh Facebook ads: a schedule that fits your budget
Ads wear out at different speeds. A starting refresh schedule by monthly budget, what makes ads tire sooner, three kinds of refresh and a short weekly routine.
- Glossary
Creative velocity
Creative velocity is the rate at which a brand makes and launches genuinely new ads, usually counted as new creatives per week. It measures the production side of testing: how quickly fresh ideas reach the ad account to replace the ones that have stopped working.
Written by the BulkAds team. Published , last checked . Meta changes its tools often — if something here no longer matches Ads Manager, tell us at contact@bulkadsai.com.
BulkAds AI is an independent product and is not affiliated with, endorsed or sponsored by Meta Platforms, Inc. Facebook and Instagram are trademarks of Meta Platforms, Inc.