Meta Ad Library Sort by Impressions: What Changed and How to Use It

What changed, in short

The Meta Ad Library sort by impressions is the biggest change to the free tool in years. For most of its life the library could only list a brand's ads by date, which pushed everyone toward the same flawed proxy: the oldest ad must be the winner. Now there is a ranking based on delivery.

But a lot of what is written about the update goes further than the evidence. This article separates what is confirmed from what is claimed, then shows how to use the sort to find ads that are scaling, and why impressions and estimated spend tell you different things. Every fact below links to its source.

What the 2026 Meta Ad Library update changed

The Meta Ad Library low impression count label

PPC Land reported on December 24, 2025 that the library had added two things two days earlier: WhatsApp as a filterable platform, and a "Low impression count" badge on ads that have accumulated fewer than 100 impressions since launch. The changes were first documented by a marketer on LinkedIn; Meta did not publish an announcement. The threshold is low on purpose. It doesn't separate winners from losers. It separates ads that got any delivery from ads that got almost none.

The impressions sort

At the end of January 2026, Coinis reported that Meta had "silently rolled out" a way to order ads by impressions. Later hands-on guides describe the control the same way: a Sort by menu, separate from the filter panel, with exactly two options, "Impressions: high to low" and "Most recent". Moda calls it "the only ranking the library gives you" in a review dated September 2, 2026.

Two caveats from the people who tested it. AdDogs found in July 2026 that impression-sorted results "skewed heavily toward ads carrying EU transparency labels" and advises treating the order as "directional rather than exact". And the same review notes the sort is undocumented: Meta hasn't said whether it reflects lifetime delivery, recent delivery, or which countries count.

What is not confirmed: impression bands on every ad

Several guides published in 2026 say every ad card, commercial ones included, now shows an impression range from under 1K to 1M+, and that you can filter by it. We could not confirm this. The two hands-on reviews above say the opposite: AdDogs puts it as "you can sort by impressions, you can't see the number, and you can't filter by it", and Moda says impression counts "stay hidden for commercial ads". Meta's own documentation still describes impression data as available for ads about social issues, elections or politics, and the API reference for archived ads marks impressions as available only for those ads.

One likely source of confusion: the filter panel has an "Impressions by date" control. It filters by when an ad delivered, not by how much. If you read a guide that says you can filter by impression volume, check whether it is describing that date filter.

How to use the Meta Ad Library sort by impressions to find scaling ads

The sort is most useful inside one advertiser, where it answers a question the library could never answer before: of this brand's 40 live ads, which ones is Meta actually delivering?

1. Search the advertiser, not a keyword. Open the brand's Page in the library, set the country you care about and filter to active ads. Moda's advice matches ours: compare ads within an advertiser, not across a keyword search, where brands of every size are mixed together.
2. Sort by "Impressions: high to low". Note the top ten. These are the ads with the most delivery behind them.
3. Read the start date next to each. A top-ten ad that started in the last 60 days is being scaled now. A top-ten ad that started a year ago is a control the brand keeps feeding. Both matter; the first one is news.
4. Check versions. "This ad has multiple versions" on a high-delivery ad means the brand is spending effort on it as well as budget. Our post on ad variants as a scaling signal explains why versions are hard to fake.
5. Skip the low impression count ads. They are tests that never got a budget.
6. Switch the country to the US and back. Because the order appears to lean toward EU delivery, a brand that sells mainly in the US can look different from one country to the next. If an ad ranks high in both, it is a safe pick.
7. Repeat next week. The sort gives no numbers, so the signal is movement: an ad that climbs from tenth to second in a week is being scaled.

Our Meta Ad Library guide covers the rest of the filters, and what "active" really means covers the start date caveats in step 3.

The combination is what works: a high-delivery ad that is new and has several versions is the pattern to look for. Two examples from our board:

Why impressions and estimated spend say different things

The sort ranks delivery. A spend figure ranks budget. They usually point the same way, but not in proportion, and the gap is where research goes wrong.

Spend is impressions times price, and the price moves. Spend equals impressions times the cost per thousand impressions (CPM). CPM changes with placement, country, audience, season and the campaign's objective. As plain arithmetic: one million impressions at a $10 CPM is $10,000; the same million at $30 is $30,000. Two ads side by side in the impressions sort can sit three times apart in budget.

Cheap delivery climbs the sort. An ad delivered mostly in lower-cost placements or countries, or bought on a reach objective, collects impressions faster per dollar than one optimized for purchases in the US. If the sort really does lean toward EU delivery, that effect is built in.

The sort only works inside one result set. It tells you which of a brand's ads gets the most delivery. It can't tell you whether that brand's best ad is bigger than another brand's, because there is no number to compare. An estimated spend puts a dollar figure on every ad, so you can rank across brands and niches.

Neither one is profit. A brand can push delivery on an ad that loses money for a while. Duration and versions are the check: brands don't keep paying for, or re-cutting, an ad for months unless it earns its keep.

The board shows why a ranking signal of any kind beats the old date proxy:

Those numbers come from our October 9, 2026 snapshot of 2,443 winning ads from 749 brands. If the money is that concentrated, sorting by start date sends you to the wrong ad most of the time, and a delivery sort is a real improvement. It is still a rank, not a budget. How a spend estimate is built and where it breaks is in estimating competitor ad spend and our methodology.

Where Ad Radar fits

Ad Radar is our product. It adds what the sort leaves out: an estimated spend on every ad, labeled as an estimate, and a board that holds only Meta ads with $50K or more of it, tagged on ten dimensions. You can follow a brand by pasting its Ad Library link and get an alert when one of its ads becomes a winner. It covers Meta only. If you would rather have a large live library without spend figures, GetHookd is the closer fit; Ad Radar vs GetHookd compares the two. For the full set of free and paid signals, start with how to find winning Facebook ads.

Figures marked as estimated spend come from Ad Radar's model of engagement on public Meta Ad Library ads. They are estimates, labeled as such, and are best used to rank ads against each other. How the board is built.