Before you start
- A web browser - no Meta login required for basic searches
- A list of 5-12 competitor Facebook Page names
- A folder structure ready for screenshot/asset capture
- 30-45 minutes for the first audit, then 10-15 minutes weekly
The playbook
8 steps
Open the canonical Ad Library URL
Go to **facebook.com/ads/library** directly. Don't click through Google search results - some of them route through redirects that lose filter state. Bookmark this URL; you'll return to it weekly.
Expected outcome
You land on the Ad Library home with the country and category filters at the top.
Set the country filter explicitly
Meta defaults the country to whichever region it thinks you're in. This is wrong roughly 30% of the time (VPNs, datacenter IPs, mobile data). Set it manually to the country you're auditing. If a brand runs in multiple countries, run the audit separately per country - the library partitions by region, not language.
Expected outcome
Country filter shows your target country, not 'United States (default)'.
Switch ad category to 'All ads'
The default category is 'Issues, elections, or politics' (the regulatory backbone). For commercial research switch to 'All ads' - this is the single most-missed configuration step in the whole tool.
Expected outcome
The category dropdown reads 'All ads' and your subsequent searches return commercial ad results.
Search by advertiser name (not keyword)
Type the brand's exact Facebook Page name. Keyword search returns every ad whose copy contains the term - which on Meta means thousands of unrelated affiliate listings. Page-name search anchors on a specific advertiser. If you don't know the canonical Page handle, open the brand's Facebook profile and check the URL slug.
# Example: Nike's page is at facebook.com/nike # So search for: nike # Not: 'shoes' or 'running shoes'Expected outcome
Search results show only ads from that specific advertiser - typically 10-100 ads per established brand.
Apply the right filters
Set **Active status** to 'Active' (focuses on what's running right now). Set **Media type** to your research target - 'Video' for hook research, 'Image' for static research, 'Carousel' for product-feed work. Don't mix media types - the structural patterns that win on video don't win on static.
Expected outcome
Result set narrows to active ads of a single media type from your target advertiser.
Read the start dates as your fatigue signal
Every active ad shows 'Started running on [date]'. Sort or scan by date. **Ads alive >60 days are the evergreen winners** - they survived Meta's auto-rotation and kept performing. **Ads from the last 14 days are tests** - most won't make it. Cluster by start date and you'll see the brand's creative cadence clearly.
Expected outcome
You can identify which 3-5 ads are the brand's evergreen winners worth deep teardown.
Click 'See ad details' for the structural reveal
Click the 'See ad details' button on any ad. The expanded view shows platforms (Feed/Reels/Stories/Marketplace/Audience Network), languages, and - critically - **all the creative variants of the same ad**. Variants tell you exactly which axes the brand is testing: hook line vs CTA, square vs vertical, frame 1 vs frame 3.
Expected outcome
You can see how the brand is testing variations of each winning concept.
Save with structured tags - not screenshots only
Capture each ad with: video/static file + start date + hook archetype + format + angle. Folder structure: `competitor-name / YYYY-MM-DD-started / hook-archetype.mp4`. Without the tag structure you'll have 50 ads in a week and no usable index later. The library is a swipe-file *input*; the tagged corpus is what compounds.
Expected outcome
Each saved ad is structured and searchable - the first week's audit feeds the second week's compound knowledge.
Shuttergen
Skip the manual audit. Get a tagged corpus weekly.
Shuttergen runs the audit for you - every competitor, every week, with hook archetype and time-on-platform tagged automatically. The eight steps above become a single dashboard.
Pitfalls
What goes wrong
Searching by keyword instead of Page name
Keyword search returns thousands of irrelevant affiliate listings, drop-shippers, and unrelated brands. Always search by exact Page name.
Forgetting the country filter
If you're auditing a UK brand and the country is set to United States, results return empty. The library partitions by region of delivery, not by brand HQ.
Treating screenshot piles as a swipe file
An untagged folder of 50 ad screenshots is a screenshot pile, not a swipe file. Without hook/format/angle tags you can't search the corpus later. Tag at capture time, not afterwards.
Running the audit once and never again
Compound intelligence comes from the second and third pass. One-time deep audits surface the current state; weekly sweeps surface the evolution. The evolution is where the strategic insight lives.
Assuming the library tells you what's working
The library shows what's RUNNING, not what's WORKING. Use time-on-platform (>60d) as the closest free proxy for performance, but don't conflate the two. A brand can run a losing ad for 90 days if they're just bad at killing tests.
Limits
When this playbook won't work
- Brands that run ads through agency-managed Pages with different names than the consumer brand
- Niches dominated by political/issue advertising where commercial-ad coverage is thin
- Categories with high creator-led affiliate spend (the library shows the brand but not the affiliate variants)
- Brands running exclusively on TikTok or LinkedIn - those have separate ad libraries
- Inactive-ad research beyond ~90 days back - older inactive ads vanish from public view
The weekly cadence that compounds
The eight steps above describe the first audit. The bigger value comes from the weekly sweep - 15 minutes per competitor, every week, for the duration of the year. Most teams skip this and lose 80% of the library's strategic value.
Monday 15-minute sweep, per competitor. Search advertiser, filter Active + Video + Last 7 days. Note any new ads. Capture them with the standard tag structure. Compare against last week's set - which ads got killed, which got new variants. The deltas are the signal.
Friday 60-minute analysis. Walk through the new captures from the week. Identify hook patterns. Cross-reference with your own ads' performance. Update your team's creative brief based on what's shifting in the competitive set.
Quarterly deep audit. Once a quarter, walk through every active ad from each competitor. Identify the evergreen winners (>60 days), the failed tests (paused after <14 days), and the patterns that emerged. Write a one-page state-of-the-competition memo. This is the doc that informs the next quarter's creative strategy.
Teams that follow this cadence for 12 months consistently outperform their competitive set on creative; teams that audit once and never again don't. The library doesn't compound automatically - your process does.
Skip the manual audit. Get a tagged corpus weekly. Shuttergen runs the audit for you - every competitor, every week, with hook archetype and time-on-platform tagged automatically. The eight steps above become a single dashboard.
What to do with the data once you have it
Captured ads are not the goal. Generating creative against the patterns is the goal. Three downstream uses for a well-tagged corpus:
Brief inputs. Pull 5-10 reference ads from the corpus into your next creative brief. Reference ads compress 500 words of description into 5 links - the editor or AI generator inherits the visual taste without reverse-engineering it from prose.
Hook archetype refresh. When your team starts producing same-y output, run a corpus sweep across your top hook archetypes. The corpus shows which archetypes your competitors are also leaning on (saturation) and which are underused (whitespace). Whitespace is where you ship next.
Performance forensics. When you ship an ad that underperforms, look up the same hook archetype in the corpus. Is the archetype itself dying in your category? Or is your specific execution off? The corpus distinguishes the two.
Internal: Anatomy of a good Meta Ad Library audit walks through a complete audit in real time; the Foreplay deep dive covers the leading tool for automating the capture-and-tag step.
FAQ
Frequently asked
Do I need a Facebook account to use the Ad Library?
Can I download videos from the Facebook Ad Library?
Why are my Facebook Ad Library search results empty?
How far back does the Facebook Ad Library go?
Is there an API for the Facebook Ad Library?
How long does a good Facebook Ad Library audit take?
Related
Keep reading
Skip the manual audit. Get a tagged corpus weekly.
Shuttergen runs the audit for you - every competitor, every week, with hook archetype and time-on-platform tagged automatically. The eight steps above become a single dashboard.