Meta Ads

How many new ads does your store need each month on Meta?

Every ad on Meta wears out, and you can’t tell in advance which new one will replace it. So the useful answer to “how many ads do I need?” is a monthly number, sized to your spend and your hit rate. Here are three ways to work it out, a worked example at three budgets, and how to keep the cost of testing down.

By the Dokiri Studio team · November 6, 2026 · 7 min read

In this article

Why you need a steady flow, not a big launch

Bluesense Digital, an e-commerce agency whose presenter publishes long Meta Ads breakdowns on YouTube, models each ad as having a finite lifetime spend: it ramps up, plateaus, then fades, faster if you push it hard (video). Growth therefore needs new ads arriving before the current ones run out.

The second reason is that no one picks winners reliably in advance. His estimate is that 5 to 15% of new ads become winners, so plan on roughly 20 ads per winner (video). These are one agency’s numbers, not Meta’s, but the pattern is familiar to anyone who buys media: most ads don’t work, and you can’t tell which by looking.

Three ways to size your monthly volume

Bluesense uses three models and checks them against each other (video). Treat all three as rules of thumb, not laws.

  1. 1

    One new ad per $1,000 of monthly spend

    The quick version: $20,000 a month means about 20 new ads. He stated it in Australian dollars and says it holds roughly in US dollars, with diminishing returns at very high spend.

  2. 2

    Work back from revenue

    Average lifetime spend per ad × ROAS = revenue per ad; divide your new-customer revenue target by it. If an average ad spends $1,000 at 3× ROAS, it brings $3,000, so a $60,000 target needs about 20 ads that reach that level. Ads can take months to earn out, so treat it as a planning figure.

  3. 3

    Replace what you lose

    Winners that fade each month ÷ hit rate. If two winners fade a month and one ad in ten becomes a winner, you need 20 new ads just to stand still.

When the three disagree, the gap tells you something. If the revenue route asks for far more ads than the spend rule, your target may be ahead of your budget. Check the ROAS you plug in against your break-even point first, with the break-even ROAS calculator.

Worked example: $5k, $20k and $50k a month

Here is the spend rule applied to three stores, with two more of his guides: put at least 10 ads into a concept before you judge it, and expect a 5 to 15% hit rate.

Monthly spendNew ads a monthConcepts you can judgeWinners to expectWhat it looks like
$5kAbout 5One every two monthsOne every 1 to 4 monthsFounder-made content and cheap formats; test angles, not polish
$20kAbout 20About 21 to 3Two concepts a month, 10 ads each, mixing openings and formats
$50kAbout 50About 52 to 7Five concepts a month; proven winners rebuilt in better production
A month at $20k, written as a plan (product: a cabin bag)
Concept A: people who fly cabin-bag only and hate paying for luggage. Pain angle, no offer. 10 ads: 5 openings × 2 formats (problem → solution, lifestyle).

Concept B: people buying for a partner who travels for work. Gift angle, gift wrapping included. 10 ads: 5 openings × 2 formats (gift idea, 3 reasons to buy).

Ongoing: new openings on last month’s winner while it still earns spend.

Close: any concept that didn’t earn spend by the end of its test window.

At $5k the arithmetic is uncomfortable: the spend rule gives five ads, but a fair test wants ten per concept. The honest answers are fewer concepts, longer windows (he suggests 10 to 14 days for small budgets), and the cheapest ad that still carries the idea.

At $50k the problem flips. You can afford volume, so the risk becomes 50 ads on one idea. He describes a brand that launched 200 ads a month on a single concept: volume without diversity. Spread the count across concepts that are genuinely different.

Set a quality bar, then maximise volume above it

Volume only helps above a minimum standard. Bluesense’s approach is to define a bar, roughly 7 out of 10, then make as many ads as possible that clear it, because even experienced strategists can’t say which of five good ads will win. He also describes accounts that launched thousands of low-effort ads a month and stayed unprofitable. Junk volume doesn’t count.

A practical bar for a product video ad:

  • The first seconds make clear who it’s for, even without sound.
  • The person on screen looks like your buyer.
  • Every claim is on your product page.
  • Captions are legible on a phone and clear of the platform’s buttons.
  • It ends with a clear next step or offer.
  • Nothing shows the product doing what it can’t.

Our creative QA checklist goes further.

Test cheaply, then upgrade the winners

Production cost decides how many concepts you can afford to try. Bluesense frames it as profit per ad: revenue minus media spend minus production. In his example, an ad that brings $1,000 at 2× ROAS leaves $480 if it cost $20 to make and $300 if it cost $200 (video). In his experience, static images are the cheapest way to test an angle, and he turns the static winners into video afterwards.

  1. 1

    Test the angle cheaply

    A static, or a simple 15 to 30-second video. The question is whether the angle and persona earn spend, not whether the edit is beautiful.

  2. 2

    Keep what earns spend

    Judge on spend and cost per purchase over a fair window, not the first day. See account structure for how to read a test.

  3. 3

    Rebuild the winner in more formats

    Same persona, angle and offer, told as a problem → solution, a lifestyle piece, a before / after.

  4. 4

    Upgrade production where it shows

    Smoother motion, sharper product shots, a longer cut for colder audiences.

  5. 5

    Iterate the opening

    New hooks on a proven body are the cheapest way to extend a winner’s life. The hook generator gives eight to start from.

What this volume costs in Dokiri

Each video’s price in credits is shown before you pay: a typical 30-second Standard video costs about 27 credits, and longer videos or more shots cost more. Writing scripts is free, so you can drop weak concepts before paying. As of November 2026:

  • About 5 videos a month (a $5k store): Pro, 190 credits a month, covers about seven 30-second Standard videos; Creator, 70 credits, about two.
  • About 20 (a $20k store): Studio, 625 credits, about 23.
  • About 50 (a $50k store): the custom plan, from 50 ads a month, its price set on a call.

Premium costs about six times as much, so keep it for winners. Unused plan credits roll over. A batch of 5 or 10 concepts varies format, person and opening for you, and its ZIP names each file for Meta. Dokiri doesn’t make statics, run campaigns or tell you which ad won: you read that in Ads Manager. Full details on plans.

Frequently asked questions

Does Meta say how many ads to run?

Not as a monthly number. Its Help Center says testing new creative matters but that too many ads at once slows learning, and it caps ads running or in review per Page: 250 for a Page whose highest-spending month is under $100,000, as of November 2026.

Is one new ad per $1,000 of spend a rule?

It’s one agency’s rule of thumb and a good starting point. Check it against your revenue target and against how fast your winners fade.

Do hook variations count as new ads?

As ads, yes; as concepts, no. Ten openings on one body test the openings of one idea. Count concepts separately from ads.

What if I can’t afford ten ads per concept?

Test fewer concepts, give each a longer window, and make the cheapest version that still carries the idea. A concept judged on three ads hasn’t really been tested.

Should every ad be a video?

No. Statics are often the cheapest way to test an angle, and in Bluesense’s experience video tends to convert better once you know what to say. Most stores need both.

Try it

Tools in this article

Turn your product into ads today

Drop a product photo or a store link: the scripts are written for free, and you pay only for the videos you make.

Make my first ad

Keep reading

All articles