A rival's Meta Ad Library page lists 240 active ads. Your account has 6 live. The number gets screenshotted into a meeting, somebody says we are being out-tested 40 to 1, and by Friday there is a plan to quadruple creative output.

I have made that call and paid for it. Those 240 entries were not 240 decisions.

Short answer: The number of ads a competitor runs, as the Meta Ad Library reports it, counts listing entries rather than distinct creative ideas. Dynamic creative permutations, per-placement builds, language copies and per-country listings each occupy their own row, so group the listing into concepts by hook and offer before you compare it against anything.

The takeaways

One concept can occupy a dozen rows. A single hook built for 1:1, 4:5 and 9:16, shipped in 3 languages, lists as 9 entries before anyone has tested a second idea.
The comparison only works if both sides get the same rule. Collapse your own account into concepts too and a 240-to-6 gap often closes to something like 11-to-6.
Run duration is the quality filter. An ad still live after 30+ days is the durable signal, while a wall of week-old entries is usually one campaign fanned out.

Why does an ad library listing show more ads than anyone is testing?

Because the library reports the unit the ad account was built in, and that unit is rarely the idea. Meta lists an entry per ad object. One concept produced in 3 aspect ratios becomes 3 objects. Ship it in German, French and Spanish and it becomes 9. Change the country filter and the same builds appear again under a different market.

Dynamic creative adds a second layer in the opposite direction. An entry carrying the multiple-versions label is one object holding a set of assets Meta recombines at delivery, so that row understates the permutation count while its neighbors overstate the idea count. Both errors sit in the same list, pointing opposite ways.

None of this is a competitor bragging. It is a campaign structure choice, and you can read it off the listing: same image, different headline, same start date, card after card.

How do you turn an ad list into a concept count?

Sort the listing by start date and walk it, grouping as you go. Two entries belong to the same concept when the hook and the offer are the same and only the wrapper moved: a crop, a caption, a language, a placement, a color. Each group gets one line in a sheet.

You need 4 fields per group. The hook in your own words. The offer. The earliest start date. And how many entries the group spans.

That last field is the one that pays. A group spanning 14 entries says the competitor invests production depth per idea, usually because a designer builds every winner out across placements. A group of 1 says they are throwing spaghetti and watching.

An afternoon gets you through one competitor at this level. That is the honest cost, and it is why most people quote the raw number instead.

Does a higher ad count mean a competitor is testing faster?

On its own, no. Testing velocity is new concepts per unit of time, while the raw count is a stock: everything live today, including ideas that have been running since March.

Two accounts can both show 240 entries and mean opposite things. One has 11 concepts, 3 of them past 90 days, the rest built out across placements and markets. That is a settled account defending a control. The other has 60 concepts, none older than 3 weeks, which is an account still hunting for a hook.

Flow is what you were trying to measure, and flow only shows up when you watch the same page over several weeks and note what disappeared. That reading has its own method, which I covered in how often competitors change their ads.

In a meeting the raw number turns a structural difference into a performance verdict, and the verdict tends to arrive as a budget request for creative volume nobody has shown you need.

What should you compare instead of the raw number?

Two numbers, both computed the same way on both accounts. Concepts launched in the last 30 days, which is the closest honest read on testing velocity. And concepts still running past 30 days, which is the read on how many of those attempts survived.

The ratio between them carries the information. 20 new concepts and 1 survivor is an account burning creative to find a hook. 4 new and 3 surviving is an account that knows what it sells and is defending it. Neither shape is better in the abstract, and which one you want depends on whether you already have a control worth protecting.

Then run the same rule over your own 6 ads. In my experience that step reverses the panic more often than it confirms it, because 6 ads in a small account are usually 6 concepts, while 240 entries in a large one are frequently 11.

Where does a normalized concept count stop being useful?

At the comparison between two different businesses. Concept counts are shaped by catalog size, funnel length and how many markets a brand sells into, so a 40-SKU retailer and a single-product subscription do not sit on the same scale. No amount of grouping fixes that.

Longitudinally the number holds up well: same competitor, same rule, measured monthly. Movement in their concept count tells you something about them. The absolute level against yours mostly tells you the two business models are shaped differently.

And the library still hides the half that matters. It shows what is live, never what earned money, so a competitor keeping a weak concept alive for 60 days is reporting their review process to you. Which of their ads deserves study at all is a separate job, and I walked through it in how to find competitor winning ads.

The habit worth keeping

Before you quote a competitor's ad count anywhere, count your own the same way. It takes 10 minutes and it removes most of the fear the screenshot created.

Concept-level grouping is the part that refuses to scale by hand, and doing it monthly across 3 competitors is a day nobody wants to spend twice. That is the problem I built the competitor side of Adscalr around: it pulls the Meta, TikTok and Google Ads Transparency libraries into one normalized dataset refreshed daily, decodes each static ad into structured fields so grouping by hook and offer becomes a query, and flags the 30+ day survivors separately from the week-old noise. The competitor intelligence pillar covers how that feeds the rest of the loop.

How Many Ads Do Competitors Run?