How Many Facebook Ads Should You Run
See how many Meta ads real Shopify stores run at your revenue level. Live counts from 1,436 stores across six revenue tiers, with a screenshot behind every figure.
Every answer to this question is a number with no source behind it. Run 3 to 5 ads. Test 20 creatives a month. Run 2 to 3 per ad set. The numbers disagree with each other by a factor of ten and none of the pages carrying them will tell you where the number came from.
There is a second problem underneath that one, and it is the reason the first problem never gets fixed. Almost every framework keys its answer to your daily budget. Daily budget is not something anyone can look up. It is not in the Meta Ad Library, it is not in any spy tool, and it is not in Brandsearch either. So a framework built on that axis can drift a long way from reality without anyone being able to check it.
What can be looked up is how many Meta ads a store has live right now. So we counted. On 12 September 2026 we measured 1,436 Shopify DTC stores across six revenue tiers, and 1,182 of them were running at least one Meta ad.
The six revenue tiers, measured
Ordered by monthly revenue. The median and the middle half count only the stores that were running at least one Meta ad on the day of the pull. Tier sample sizes, smallest band first, are 111, 273, 268, 274, 285 and 225 stores.
| # | Monthly revenue | Running Meta ads | Median live ads | Middle half |
|---|---|---|---|---|
| 1 | $10k to $50k | 62% | 25 | 5 to 65 |
| 2 | $50k to $100k | 73% | 57 | 11 to 178 |
| 3 | $100k to $500k | 79% | 121 | 45 to 332 |
| 4 | $500k to $1M | 84% | 165 | 58 to 398 |
| 5 | $1M to $5M | 91% | 311 | 123 to 769 |
| 6 | $5M and up | 94% | 391 | 138 to 1,039 |
How to read the numbers
These are live ad counts, not budgets. The figure is how many Meta ads that store had running on 12 September 2026. It says nothing directly about what those ads cost, and the rest of this page is fairly blunt about why you should stop trying to infer one from the other.
The median is a description, not a target. Look at the width of the middle half before you take the median seriously. In the $1M to $5M tier the middle half runs from 123 to 769 ads, so two stores of the same size routinely differ by six times. The spread inside a tier is larger than the gap between tiers, which means "how much do you make" explains far less about ad volume than every framework on this subject assumes.
An ad is not an idea. A large share of these counts is one creative duplicated across placements, audiences and catalogue variants. That is the single most important correction on this page and it has its own section below, with the numbers.
Euro figures are the disclosed slice, and rank badges are worldwide. EU and UK rules force advertisers to publish spend and reach on some ads. Nothing forces disclosure elsewhere. So a store selling almost entirely into the US can run hundreds of ads and show a tiny euro figure, and that is a disclosure artefact rather than a small budget. Rank badges work the other way: they are computed across everything a brand runs in every market, so a worldwide number sits next to a regional one on the same card.
Revenue itself is an estimated band, which is why the tiers here are wide.
1. retrosportsfactory.com
This UK store sells personalised retro sports shirt gifts, and it sits almost exactly on the median for the smallest tier we measured: 24 live ads out of 109 that Brandsearch holds for it. Traffic is 79.2k a month. Sixty-two percent of stores in this band run any Meta ads at all, so simply being here puts a store ahead of a third of its peers.
The interesting figure is the one on the right of the ads card: €78.3k of disclosed EU spend against 24 live ads. Hold that number, because a store forty times its size appears further down this page with a disclosed figure six times smaller. The store is 70% UK traffic, and UK disclosure rules are why almost all of its spend is visible.
2. pointnopointstudio.com
Point No Point Studio makes salt and pepper engagement rings from recycled materials and has been in the index since 2016. It runs 58 live ads out of 133 held, one ad off its tier's median of 57, on 87.6k monthly visits.
Watch the blue line rather than the big number. Over six months this store's live count moves between 32 and 70 and ends at 58. There is no single correct ad count even for one store in one quarter, which is worth remembering the next time a guide gives you one number and no range.
3. sowears.net
Sowears sells dresses, co-ords and modest fashion, and at 121 live ads out of 708 it lands exactly on the median for the $100k to $500k band. Traffic is 406.9k a month and the ads run across 8 Facebook pages, which is the first sign on this list of an operation structured rather than improvised.
The 708 is worth reading carefully. It is the Meta ads Brandsearch currently holds for the store, not a lifetime tally, because old creative ages out of the index. The ratio of live to held is a rough read on how fast a store cycles creative: Sowears keeps about one ad in six alive.
Growth in ad count is its own signal
Sowears is down 8% on its live ad count over six months while its revenue estimate has been recovering, which is the opposite of what the "more revenue means more ads" story predicts. The next store on the list is the clean version of the pattern the story wants.
4. ion8.co.uk
Ion8 sells leak-proof water bottles for adults and kids. It runs 155 live ads out of 421 held, on 513.3k monthly visits, with €632k of disclosed EU spend. Its median tier peer runs 165, so this is an ordinary volume for its size.
What is not ordinary is the slope. The live ad count has gone from 67 to 155 in six months, up 158%, and the revenue estimate has climbed alongside it. This is the shape people are actually looking for when they ask how many ads to run: not a number, but a direction and a rate. A snapshot of one competitor cannot show you that. Six months of daily history can.
Check your own live ad count against the stores in your revenue band
5. fiido.com
Fiido sells ebikes into Germany, Italy, the Netherlands and Poland. It runs 363 live ads out of 8.8k held across 42 Facebook pages, and its disclosed EU spend reads €4.3M.
Two things here. The 42 pages explain the 8.8k: a store running ads from dozens of pages accumulates held creative fast, and the live count stays a small fraction of it. And because Fiido sells into the EU, nearly all of its spend falls under disclosure rules, so the euro figure is close to the real thing. Compare it with the next entry.
6. fashionnova.com
Fashion Nova is the largest store on this list by a wide margin, at roughly $7M a day in estimated revenue and 39.1M monthly visits. It runs 343 live ads out of 519 held, which is below the median of 391 for its own tier and within twenty ads of Fiido, a store several times smaller.
Its six-month ad line is the most volatile on the page: 73, then 684, then back to 343. The same brand answered "how many Facebook ads should I run" with 73 and with 684 inside half a year. Any framework that hands you a fixed number for a revenue level is describing one frame of that line and calling it the rule. Read campaign structures presented as a template with the same suspicion.
Ad count is not a budget proxy
This is the correction that matters most, because the advice it replaces is everywhere: use a competitor's ad count as a stand-in for their budget.
Put the two large stores side by side. Fiido runs 363 live ads and discloses €4.3M. Fashion Nova runs 343 live ads and discloses €13k. Twenty ads apart, and a factor of roughly three hundred apart on the euro figure. Neither store is spending three hundred times the other. Fiido sells into the EU, where disclosure is mandatory, and Fashion Nova sells 79% into the US, where it is not.
The same trap works in reverse at the small end. retrosportsfactory.com runs 24 ads and discloses €78.3k, six times Fashion Nova's disclosed figure on one fourteenth of the ad count.
So read the euro figure as what it is, a window onto the EU and UK slice of a budget, and never as the budget. It is extremely useful for that, and it is worthless as a scaling multiplier.
Stop reading about winners. Find them yourself.
Search 6.5M+ brands, their ads, revenue, and products — all in one place.
Try Brandsearch freeCount concepts, not ads
Here is why the raw counts in the lead table run so far above every published recommendation, and why chasing them would be a mistake.
Comfrt sells hoodies, blankets and loungewear. The Hooks view reads its ads by opening line and reports 180 opening lines across 1,370 of the ads we scanned, with a further 146 catalogue ads whose text Meta fills in per product. So roughly 1,370 ads resolve to 180 ideas, about seven and a half ads per idea.
The concentration at the top is sharper still. The single line "The Most Viral Hoodie Brand Online!" carries 155 ads on its own, more than the entire live ad count of four of the six stores in the lead table. The next four lines carry 81, 77, 76 and 53. A brand at this scale is not writing hundreds of ideas. It is writing a handful and duplicating the winners hard.
Fashion Nova gives the same shape from a different catalogue: 211 opening lines across 1,382 scanned ads, plus 308 catalogue ads. The top line, "New to Fashion Nova?", carries 120 ads. Then "Sun, skin, and new Nova Swim." on 103, and "Still looking for swim that actually hits right?" on 77.
This is the answer to the apparent contradiction between a guide telling you to run 5 ads and a table showing a median of 391. Both are describing real things at different units. If you count ideas, the recommendation and the measurement stop fighting: even a store running well over a thousand ads is working from a couple of hundred lines, and leaning on five.
Pull the hook list for any brand in your niche and see how many ads sit behind each line
When more ads actually hurt
Duplication is cheap. Splitting a budget is not, and this is where the one genuine mechanism in this topic lives.
Meta's documented learning-phase threshold is roughly 50 conversion events per ad set per week. Note the unit. It is per ad set, not per ad, and several versions of this advice get that wrong, which matters because the entire framework is derived from it. Spreading one budget across many ad sets is what starves learning. Adding creative inside an ad set that is already converting does not carry the same cost, which is exactly why duplication at the top of this page is a viable strategy rather than a wasteful one.
The arithmetic is worth doing once, plainly, as arithmetic rather than as data. At a $25 cost per acquisition, an ad set needs about $1,250 of weekly spend to reach 50 conversions, or roughly $180 a day. Below that the ad set is learning slowly whatever you do, and the fix is fewer ad sets rather than fewer creatives. That is the floor arithmetic sets. The other floor is competitive rather than statistical, and the people you bid against are the ones who set it.
The second failure is concentration you cannot see. The Hooks numbers above are the measurable form of it: when one line carries 155 of 1,370 ads, the rest of the account is a long tail, and the useful question about that tail is whether it is testing or coasting. Brandsearch marks each ad with a phase, Scaling, Evergreen, Fading or not yet measured, so the tail can be read rather than guessed at.
See which of a competitor's ads are scaling and which are fading
What the six have in common
Three things fall out of the table and the six stores, and none of them is a target ad count.
Ad volume rises with revenue, but weakly. The median climbs from 25 to 391 across the six tiers, a fifteen-fold rise across a revenue range of several thousand fold. And the share of stores advertising at all climbs steadily, from 62% in the smallest band to 94% in the largest, which is a more reliable pattern than the counts themselves.
Volume is unstable within a single store. Every six-month line on this page moves: Fashion Nova between 73 and 684, ion8 from 67 to 155, Point No Point between 32 and 70. We did not measure a seasonal effect and will not claim one, but no store on this list held a steady number for two consecutive months.
The live count is always a fraction of the held count. 24 of 109, 58 of 133, 121 of 708, 155 of 421, 343 of 519, 363 of 8.8k. Creative turnover is the constant, not the headcount.
Every one of those six-month lines exists because the count was recorded daily for months before anyone asked the question. A single snapshot of a competitor cannot produce one.
Track a competitor and watch the count move day by day
How to run this for your own niche
The Brand Library card carries the two figures this page is built on. Quince shows 10.5K live Meta ads of 10.6K held at $12.3M a day and about €1.4k a day disclosed, which is a store cycling almost nothing out. Ign shows 13 of 339 at a similar revenue, which is a store that has all but stopped. That is the same revenue answering the ad-count question two opposite ways, on one screen.
That capture is the raw list, sorted by revenue and otherwise unfiltered, which is why a media site sits next to a fashion store in it. The measurement in the lead table used a filter: Shopify, brand type DTC, at least 50,000 monthly visits so the revenue estimate means something, at least three products, and the six revenue bands. Stores whose revenue estimate implied more than $25 per monthly visit were dropped as broken estimates, and regional sub-stores were dropped because they show a different ad set.
To do it yourself: filter to your niche and your revenue band, read the live and held counts on each card, then open the two or three that are moving and look at their Hooks. Save them to a swipe file tagged by revenue band. Then point Brand Analysis at your own store and run the same read on yourself, which is the step almost nobody does.
Run the same filter on your own niche
The API and MCP
Every figure in the lead table came out of one endpoint. This is the real query for the $1M to $5M row, unedited:
curl -s -H "X-API-Key: $BRANDSEARCH_API_KEY" \
"https://api.brandsearch.co/v1/brands?platform=shopify&brand_type=dtc%20brand\
&min_revenue_min=1000000&min_revenue_max=5000000\
&monthly_visits_min=50000&product_count_min=3\
&sort_by=monthly_visits&sort_order=desc&page_size=100\
&fields=id,name,niche,monthly_visits,min_revenue,last_meta_active_count,last_meta_total_count"
Change the two revenue bounds and you have the other five rows. One warning worth having: when last_meta_total_count comes back as 50001 it is a ceiling sentinel rather than a count, so drop it instead of printing it.
The same data is available over MCP, so Claude, Cursor or ChatGPT can run this count against your own niche and answer the question with your numbers rather than ours.
The bottom line
Pick a number of ideas, not a number of ads. Two to five live concepts is a real target at any size, because even a store running 1,370 ads is working from a handful of lines and duplicating the ones that hold.
Then let the ad count fall out of two things: how many ad sets you can fund past Meta's learning threshold, and how hard you duplicate what works. If you want a benchmark rather than a rule, the median live count for your revenue band is in the table at the top of this page, and the middle half beside it will tell you how little that median is worth on its own.
Questions people actually ask
How many Facebook ads should I run per ad set?
Enough that the ad set clears roughly 50 conversions a week, which is a budget question rather than a creative one. Adding creatives inside a funded ad set is cheap; adding ad sets that each fall below the threshold is what stalls learning.
How many ads do successful ecommerce brands run?
Measured across 1,436 Shopify DTC stores on 12 September 2026, the median live Meta ad count was 25 at $10k to $50k a month, 121 at $100k to $500k, and 391 above $5M. The middle half of each band is wide enough that the median is a description rather than a recommendation.
Is it better to run more ads or better ads?
Neither framing survives contact with the data, because the largest advertisers do both at once: a small number of ideas, duplicated many times. Comfrt runs 180 opening lines across 1,370 scanned ads and puts 155 of them behind one line.
How long should I leave an ad running before killing it?
That is a different question with a different answer, and it is covered in how long to run a Facebook ad before killing it. For durability specifically, ads still running past 90 days is the stricter filter.
How do I benchmark against my own niche rather than a revenue band?
Revenue bands are blunt. The niche version of this method is in how many ads to run, benchmarked by niche, and if you count creatives rather than ads, how many ad creatives to run and benchmarking against competitor active ad counts both cover that unit.
Counts in this article were measured on 12 September 2026 and every one of them is visible in the screenshot beside it.
