Meta Ad Library competitor research: the operator's playbook

The Meta Ad Library is the single most under-used research surface in paid social. Mark Zuckerberg’s transparency tool, the regulatory artefact that nobody at Meta wanted but everyone in growth marketing now relies on, lives at facebook.com/ads/library and exposes every active ad on every Page on Meta’s properties. No login wall, no auth, no rate limit on basic browsing. And yet most operator teams use it for the same thing every time — checking one competitor’s ads once a quarter when the team lead asks “what’s Klaviyo running these days?” — and miss the much larger workflow it enables. This is the operator playbook for 2026.

TL;DR

• The Ad Library is free, comprehensive (every active ad across Meta’s properties), and lets you filter by region, format, date range, and Page.
• Six high-signal workflows earn the time: angle mining, fatigue detection, format adoption, geo expansion tells, creative refresh cadence, and brief-killing.
• The tools layer — Foreplay, Motion, Atria, and the Ad Library API — turns one-off browsing into ongoing swipe files and structured datasets.
• AI ad-research agents are starting to query the Ad Library programmatically, which is going to change the competitive intelligence game across the next 12 months.
• The five competitive intelligence patterns in section six of this piece are the ones we run weekly. The rest is noise.

What the Meta Ad Library actually shows

Quick reference for anyone who hasn’t opened it lately. The Ad Library exposes every active ad currently running on Facebook, Instagram, Messenger, and Audience Network. For each ad:

• The creative (video, image, carousel, collection — whatever the format).
• The Page running the ad.
• The “started running on” date.
• For political/social-issue ads only: the spend range, impressions, and demographic breakdown.
• The countries and platforms the ad is being shown in.
• Whether the ad is part of a campaign with multiple variants (and a count of those variants).

For non-political ads — which is most of what you’ll be researching — the spend and impression data are not exposed. That’s the load-bearing limitation. You can see what’s running and how long it’s been running. You can’t see what it’s spending or how it’s performing.

The “how long it’s been running” signal is the substitute for spend data. An ad that’s been running 90+ days on a competitor’s account is almost certainly profitable for them; an ad that vanished after 7 days is almost certainly a failed test. This pattern recognition is the entire engine of Meta Ad Library competitive research.

The six workflows worth the time

1. Angle mining

The most useful workflow. Open three to five competitors in adjacent tabs. Sort each by “active for longest” (Meta exposes the start date on each ad). The ads at the top of each competitor’s list are the angles that have survived their internal kill criteria — copy hooks, value propositions, social proofs, demos that the competitor’s algorithm has decided to scale.

Write down the angle, not the creative. The creative is the rendering; the angle is the asset. “Founder-led explainer with a 7-day refund guarantee in the second beat” is an angle. “Woman in green sweater holding the product” is creative direction.

The fastest way to populate a creative brief is 30 minutes of angle mining across six competitors. We’ve seen brand-new operators ship their first profitable Meta campaigns off competitor angles found this way within their first two weeks on the account.

2. Fatigue detection

Sort the same competitors by “started running” date, newest first. Look at the velocity of new ads launched in the last 7 days vs the trailing 30. Spikes signal one of two things: a major campaign push (new product, seasonal moment) or creative fatigue (their old winners stopped scaling and they’re testing replacements).

Drop-offs signal the opposite — either they’ve found a winner and are riding it, or they’ve pulled budget from the channel entirely.

Pair fatigue detection with your own data. If you’re seeing CPM compression in your account and a competitor’s launch velocity dropped to near zero last week, the channel might be opening up for you.

3. Format adoption signals

Filter by format (video, image, carousel, collection, stories) for each major competitor. Track the format mix month over month. A format shift across multiple competitors in the same category is a signal that the category is moving — likely because one early mover found a format that scales and others are copying.

This is how we caught the move from In-Feed Reels to Stories carousels in DTC supplements at the end of 2025. Three brands in the category shifted carousel share from 5% to 20% inside three weeks. The format wasn’t new; the application to that category was.

4. Geo expansion tells

Use the country filter. If a US-focused competitor suddenly starts running ads in Germany, Spain, or the UK, you’re seeing geo expansion in real time. The ad copy gets translated, the visuals usually stay the same, the targeting is fresh.

For SaaS and consumer apps especially, geo expansion in the Ad Library shows up weeks before the official launch announcement. We’ve used this to map Taxfix’s expansion patterns across the UK, Germany, Spain, and Estonia by watching their Ad Library presence per market. It’s also how to track when an international competitor enters your home market.

5. Creative refresh cadence

For your own competitive set, log the launch dates of every new ad over a 90-day window. Compute the median time between new creative launches per competitor. The brands launching new creative at 2-3× the pace of their peers are the ones running agentic AI ad creative loops — they’ve decoupled creative production from human-creator capacity.

This is where the agentic shift starts showing up in the data. A competitor that used to launch one new ad per week now ships five. That’s not a freelancer hire; that’s a tool change.

6. Brief-killing

A specific workflow for ad agencies and consultancies. When a client briefs you with a creative idea (“we want to do an unboxing video”), search the Ad Library for the format applied to their category in the last 90 days. If 12 of their direct competitors have shipped similar ads recently — and most of them have stopped running them — that’s evidence the format is exhausted for the category. Better to find this out in the briefing call than three weeks into production.

The five competitive intelligence patterns

The workflows above are the methodology. The patterns below are what you’re looking for in the data. Run each one on every major competitor weekly.

Pattern 1: The one ad they refuse to kill. Every brand has one ad that’s been running for 6+ months. That ad is their hero asset. Read it three times. Reverse-engineer the hook, the offer, the social proof, the call to action. The hero asset reveals the brand’s actual value proposition (which is often different from the value proposition on their landing page).

Pattern 2: The seasonal pull. Q4 push, Black Friday, post-holiday return, summer slowdown. Track which competitors run hard into each season and which don’t. The brands that pull spend in Q1 are usually the ones with cash flow problems; the brands that pull in Q4 are usually the ones with inventory or margin problems.

Pattern 3: The format experiment. A competitor launches a single ad in a format they don’t usually run (an unboxing video for a brand that ships statics, a UGC explainer for a brand that ships studio production). Two outcomes: the experiment vanishes in a week (failed test, ignore), or it’s joined by 3-5 more in the same format over two weeks (winner found, the brand is scaling the format, copy the angle).

Pattern 4: The competitive callout. Some brands name competitors directly in their ads. Track which brands callout which competitors. The callout direction often signals where the threat is perceived. (Note: callouts in Meta ads are subject to platform policy review, so they tend to be soft — comparison tables, “us vs them” stats, careful language.)

Pattern 5: The audience expansion test. A direct-response brand suddenly running a brand-awareness style ad. The ad has fewer CTAs, more story, more lifestyle imagery. They’re testing whether they can buy at the top of the funnel cheaper than they’re buying mid-funnel. Watch how long that ad runs. If it survives 30 days, they found the unlock.

The tools layer

The Ad Library itself is excellent for one-off browsing and good-enough for weekly checks. When you need ongoing surveillance, structured datasets, or swipe files built into your team’s workflow, the tools layer matters.

Foreplay

Foreplay is the operator-favourite swipe-file tool. Save any ad from the Meta Ad Library (or TikTok, or YouTube) to a Foreplay board, tag it by angle and format, share boards across your team. The interface lives on top of the Library’s data; the value is the organisation layer. Pricing starts around $49/month/seat and scales with team size and storage. Foreplay is the closest thing to a default in this category — most operators we work with run it.

Motion

Motion integrates Meta and TikTok ad data with performance data from the ad accounts and creative metadata, and surfaces which creative attributes correlate with performance. It’s a creative-strategy tool that uses Ad Library data as one input among many. Strong for brands shipping 50+ creatives/week who need to learn from the volume.

Atria

Atria (formerly Glimpse) is the lighter-weight swipe-file tool. Newer, leaner, more affordable than Foreplay for early-stage brands. The feature set is smaller but covers the core swipe-file workflow well.

The Meta Ad Library API

The official Meta Ad Library API exposes the same data programmatically. Free tier with rate limits, requires Meta developer-account approval. The API is the substrate that powers the third-party tools above — and increasingly, custom internal tools that growth teams build for their own competitive intelligence needs.

AI agents querying the Library

The 2026 trend worth tracking: AI ad-research agents that query the Library programmatically, summarise patterns across competitor sets, and surface angle and format trends in operator-language reports. Superscale is one example of an agent that pulls Library data as part of its competitor research pillar, alongside other tools that focus specifically on the research layer. The first generation of these agents is workable; the second generation (which we expect in late 2026) will probably collapse the research-tool category into the broader agentic stack.

How to set up a competitor research workflow

Practical sequence we’d ship if standing this up for a new brand or agency client.

1. Define the competitive set. 6-10 brands max. Direct competitors plus 2-3 adjacent-category brands you respect.
2. Set up a Foreplay or Atria board per competitor and per angle category (hooks, demos, social proof, comparison ads, offer ads).
3. Run a 60-minute Monday morning sweep. Open each competitor in the Ad Library. Save anything new launched in the last 7 days to the right board. Note which old ads have dropped off.
4. Run a 90-minute quarterly deep dive. Sort by “longest running” per competitor. Pull the top 5 ads per brand. Write a one-page brief on each one — angle, hook, offer, social proof — and circulate to the creative team.
5. Feed the swipe file into briefs. When briefing a new creative, attach the Foreplay board for the relevant angle. The creative team works against examples, not just words.
6. Run incrementality on what you copy. Angles from competitors are hypotheses, not winners. Test them with the same rigour you’d test in-house ideas.

What the Ad Library can’t tell you

Three limits to keep in mind so you don’t overweight what you find.

No performance data. You see what’s running, not what’s working. The 6-month survival heuristic is your only substitute for spend or CPA data.

No targeting data. You don’t know who they’re showing the ad to. A competitor running an ad in your market might be running it to a tiny custom audience, not the broad targeting you’d assume.

No bid or budget data. A brand running 200 ads might be spending $200 or $200,000 — same ad count, vastly different scale. Be careful inferring competitor budget from ad volume.

For deeper attribution and budget context on what you find, pair the Library with platform-reported data, marketing mix modeling, and ROAS tracking on your own program. The Library is the qualitative signal layer; your data is the quantitative layer.

Frequently asked questions

What is the Meta Ad Library?

The Meta Ad Library is a free, public-facing transparency tool launched by Meta that exposes every active ad currently running on Facebook, Instagram, Messenger, and Audience Network. Anyone can browse it without an account at facebook.com/ads/library, filter by Page, region, format, or date, and view the creative directly.

How accurate is the Meta Ad Library?

The Library shows every active ad in real time, sourced directly from Meta’s ad-serving infrastructure. The data is accurate for what it covers — what’s running, when it started running, what format it is, what platforms it’s on. The Library does not show spend, impressions, or performance data for non-political ads.

Can you see how much a competitor is spending on Meta ads?

No, not for commercial ads. Meta only exposes spend and impression data for political and social-issue ads. For everything else, you can see how long an ad has been running (which is a useful proxy for whether it’s profitable) and the volume of variants the brand is running, but not the dollar figure.

What’s the best tool to use with the Meta Ad Library?

Foreplay is the operator-favourite swipe-file tool layered on top of the Library — it lets teams save, tag, and share ads across boards. Motion is the option if you need to correlate Library data with your own ad-account performance data. The Meta Ad Library API is free and the substrate for custom internal tooling.

How often should I check the Meta Ad Library?

For most growth programs, a 60-minute Monday morning sweep on your top 6-10 competitors, plus a 90-minute quarterly deep dive, is the right cadence. Brands shipping creative at high volume (50+/week) benefit from daily monitoring via a tool like Foreplay rather than manual Library browsing.

Are AI tools using the Meta Ad Library?

Yes. The Ad Library API and direct scraping power most modern competitor research tools. AI marketing agents — including Superscale and the broader agentic stack — are increasingly using Library data as one input into autonomous research workflows. The trend is toward agents that surface competitive patterns directly to operators rather than handing them raw data.

Related reading

• TikTok Creative Center operator playbook — the TikTok equivalent research surface.
• How to launch AI ads on Meta — what to do with the angles once you’ve found them.
• Winning hook patterns of 2026 — the angle taxonomy used inside our briefs.
• CPI benchmarks for mobile apps in 2026 — the install economics behind the ads in the Library.
• Creative strategist role in 2026 — who owns the swipe file inside the team.