Meta Ad Library now sorts by impressions. Here's what it still won't show you.

The 2026 update finally ranks ads by reach. It still leaves out spend, scripts, and brand health, the three things you actually need to copy a winner.

For years, operators scrolled the Meta Ad Library guessing which creative was actually working. Running days gave you a hint.

The number of variants gave you another.

That was it.

Meta's 2026 update changes that. You can now sort Meta Ad Library results by impression volume, and the top of the list is the ad that's reached the most people.

It's a real upgrade for anyone doing competitor research.

But the new sort fixes one blind spot and keeps four others. If you copy the top-impression ad without knowing the spend behind it, the script inside it, or the store health around it, you're still flying without the full context.

This article maps exactly what the new Meta Ad Library impression sort gives you. And where you still have to go to close the gaps.

What the new impression sort actually does

The 2026 update adds an "Impressions" sort option next to the existing "Most Recent" filter. Click it and the ad library reorders by reported impression range.

The ad with the most eyeballs goes to the top.

That's genuinely useful. You stop guessing which variant the brand is actually pushing.

You also stop wasting time on test ads that shipped yesterday and will be paused by next week. Combined with the "Running Days" display, you can finally find ads that have been live for 25+ days and distributed at scale.

One caveat: impression ranges are buckets, not exact numbers. Meta still reports ranges like 1M–2M.

You can tell the difference between a test ad and a scaler, but you can't tell the difference between two scalers.

The other thing to know: the sort only works on ads that have enough EU impressions to be reported. Any ad that never ran in the EU doesn't get an impression label at all.

If your competitor is US-only, the new sort is mostly dead weight for you.

Impressions aren't dollars. And Meta still won't show you either

Here's the honest limit of the new sort: impressions tell you reach, not spend.

A brand running a $200 CPM to a warm retargeting audience can hit huge impression numbers on a tiny budget. A cold-traffic video with a $20 CPM can burn way more money and still show a lower impression count.

If you're picking which ads to copy, spend is the signal you actually want. Impressions are a proxy.

Spend is the truth.

Meta doesn't publish spend numbers outside of political ads. Full stop.

No future update will change this, it's a policy call, not a missing feature.

The one place real Meta spend leaks out is the EU. Under the Digital Services Act, brands running in France, Germany, Italy, Spain, and the Netherlands have to report daily euro spend per ad set.

That data exists.

Meta just doesn't surface it inside the ad library UI.

Brandsearch EU Adspend pulls that DSA data directly. You open any tracked brand and see actual euros per day, per country, per campaign.

That's the difference between "this ad reached a lot of people" and "this ad is burning EUR 2,400 a day and hasn't been paused in six weeks."

That's not a feature comparison. That's a gap Meta will never close.

The script gap: You see the ad, not the hook

The second thing the new sort won't give you is the script.

Most top-impression ads on Meta in 2026 are videos. Image ads still exist, but the scale winners are almost always 15-to-60 second videos.

Sort by impressions, click into one, and you get a player, that's it.

Watching the video isn't enough for research. You need the opening hook, the problem setup, the CTA, and the structure in text form.

That's how you spot the pattern across ten winners in the same niche.

Three of the top five winners might open with a cost-of-inaction hook. Two might open with a raw product demo.

You can't see that by pressing play on one ad at a time.

Meta gives you one video at a time. It doesn't give you side-by-side transcripts.

The Brandsearch Scripts tab inside any Brand Analysis page does. Open a brand, click Scripts, and every video ad that brand has run gets transcribed with the hook isolated at the top.

You can scan 40 ads in five minutes and see the exact structure the brand is testing.

That's the workflow shift. From "I watched three of their ads" to "I've seen every video they've shipped this quarter."

Creative clusters and store health live somewhere else

The new impression sort answers the question "which ad is biggest". It says nothing about the question underneath it: is this brand actually scaling, or are they about to stall?

An ad with 5M impressions inside a brand that's lost 30% of its traffic in the last 90 days is not a brand you want to copy. They may be pushing harder to cover a leak.

Meta Ad Library won't tell you that, because it doesn't see the store.

The Brandsearch Brand Analysis Overview tab sits on top of the same ad data, but adds the rest of the data. You get a Traffic Trends chart showing monthly visitors over time, an Ad Scaling chart showing ad count by day, a revenue estimate, and a bestsellers strip.

That combination tells you things a single ad can't. A brand whose ad count tripled in the last 30 days while traffic also went up is actively scaling.

A brand whose ad count tripled while traffic went flat is burning money. A brand whose ad count dropped while traffic grew found something that works organically and pulled back.

You need all three signals to know which of those stories you're looking at. Ads, traffic, and time.

Impressions alone tell you none of them.

The 2026 workflow I actually run

Here's how the new Meta Ad Library impression sort fits into a real research session.

I open Meta Ad Library for a 10-minute scan of whatever niche I'm watching that week. Sort by impressions.

Skim the top 20. Screenshot the two or three that look interesting.

Then I switch tools. For each of those ads, I want four things Meta can't give me: real spend, the script, the creative cluster the brand is running, and store-level health.

I run each brand through Brandsearch in the same order every time:

1. Brandsearch Discovery, pull every active ad from the brand and check if the impression winner is an outlier or part of a pattern.
2. Brandsearch Brand Analysis (Overview tab), read the Traffic Trends and Ad Scaling charts to confirm the brand is actually growing.
3. Brandsearch Scripts tab, lift the hook, the problem frame, and the CTA from every video variant.
4. Brandsearch EU Adspend, for any brand running in the EU, read the actual euro-per-day number on the ad set.
5. Brandsearch Calculators, plug the spend estimate into Break-Even ROAS and see if the math on their scaling is plausible.

The whole loop takes maybe 15 minutes per brand. Without this stack, I'd be staring at impression buckets and guessing.

One note for anyone starting from zero: install the Brandsearch Chrome Extension first. It's free, it lives in your browser toolbar, and every Shopify store you land on shows instant ads, traffic, and tech stack.

When you outgrow it, every brand you researched carries straight into the full app.

Meta Ad Library stays in the workflow. It's useful for the first scan.

It just isn't enough on its own anymore.

Summary

The new Meta Ad Library impression sort is a real upgrade. Use it for the first 10 minutes of any research session to find the ads that are actually distributed at scale.

Then stop and switch. Track the money and the impact, not just the reach:

1. Brandsearch Discovery, filter to the brand and see every variant they're running.
2. Brandsearch Brand Analysis, confirm the store is actually scaling, not just advertising.
3. Brandsearch Scripts tab, pull the hooks in text form, not just the thumbnails.
4. Brandsearch EU Adspend, read real euros per day for any brand running in France, Germany, Italy, Spain, or the Netherlands.
5. Brandsearch Chrome Extension, free starting point for any store you land on in the browser.

Impressions are the start of the story. Spend, scripts, and store health are the ending.

Don't copy an ad until you've read all three.