How to Find Winning Ads in Meta Ad Library (And Legally Recreate Them)
The free creative research tool 95% of DTC founders never open — and the exact workflow for using it.
Most founders running paid ads on Meta are flying blind on creative.
They run their own ads. They check their own CTRs in Ads Manager. They occasionally screenshot a competitor's post that pops up in their personal feed. That's it. That's the entire creative research process for thousands of small DTC brands burning through ad budgets every month.
Meanwhile, sitting at facebook.com/ads/library — free, public, no login required — is a database of every single ad currently running on Meta. Searchable. Filterable. By keyword, by country, by brand, by date. With high-resolution previews of every creative. Updated in real-time.
Most founders have heard of it. Almost none use it systematically. This post fixes that.
What you'll get from this post
A repeatable workflow for finding ads that are actually winning in your category (not vanity ads, not brand-awareness ads — real performance ads), the signals that tell you a creative is working, the legal line on what you can do with what you find, and a tight handoff into recreating them for your own product.
What Meta Ad Library Actually Is
Meta launched the Ad Library in 2019 under regulatory pressure — originally to make political ads more transparent. It expanded to cover every active ad on Facebook and Instagram, regardless of category. It is, accidentally, the single best free creative research tool in the world.
What it contains:
- Every active ad running on Meta right now, globally
- Each ad's start date (so you can calculate how long it's been running)
- Multiple ad variants from the same campaign, side-by-side
- The destination URL the ad clicks through to
- Platform placement breakdown (Facebook feed, Reels, Stories, etc.)
- For political/social-issue ads: estimated spend and demographic reach
What it doesn't contain (and never will): exact spend, exact CTR, exact conversion data for commercial ads. You're inferring performance from proxy signals — which is what most of this post is about.
Step-by-Step: Your First Ad Research Session
Block 30 minutes. Open the Ad Library. Walk through these steps in order.
Open Meta Ad Library and set your country filter
Go to facebook.com/ads/library. The first dropdown is country. Set it to your target market, not necessarily your home country.
Pro tip: If you sell premium products, default to US / UK / CA / AU — these are high-AOV English-speaking markets where the best DTC brands concentrate spend. If you sell budget products, your reference pool is in Southeast Asia and Latin America (different aesthetic conventions entirely).
Choose "All ads" category
The second dropdown is category. "All ads" captures commercial DTC. "Issues, elections, or politics" is the political ad subset (much more granular spend data, but irrelevant to e-commerce research).
Pro tip: The political category has a separate, richer dataset. If you ever want to study ad copy psychology at scale, browse it. Not for creative research though.
Search by category keyword, not brand name
Most founders type a competitor's brand name and stop there. Bad. You're locked into one brand's aesthetic. Better: search by category keyword — "candle", "perfume", "skincare", "supplement", "pet food", "yoga mat" — and let the Ad Library surface brands you've never heard of.
Pro tip: The hidden gems are mid-size brands you don't know yet. They're often running smarter creative than the household names because they have to.
Filter to "Active" ads only
By default the Library shows active + inactive. Filter to active only. An inactive ad is one Meta killed — either underperforming or campaign-ended. Active ads are paying to run right now, which is your single best free performance signal.
Filter by ad format and platform
If you're researching for Instagram Reels, filter to video + Instagram. If you're researching for Facebook feed banners, filter to image + Facebook. Don't let Stories ads contaminate your feed research.
Pro tip: Static image ads are the easiest references to recreate. Video is harder to clone (more variables — pacing, transitions, audio). Start with static.
Click "See Ad Details" on the interesting ones
This opens the full ad card: start date, all active variants in that campaign, destination URL, platform placement. This is where the real signal lives.
Screenshot at full resolution
Right-click the ad image and "Open in new tab" — or "Save image as" if your browser supports it. Don't screenshot the small thumbnail. The Library serves full-resolution creatives for active ads. Get the real file.
Pro tip: Higher reference resolution = cleaner style extraction when you feed it to Rival Mode. Compressed JPEGs lose the lighting nuance that makes the recipe work.
How to Spot a Winner (vs. a Vanity Ad)
Not every ad you find in the Library is worth copying. The Library shows everything — including ads that are about to be killed for poor performance. You need to filter for ones that are actually working.
You can't see CTR. You can't see spend. But you can see these proxy signals — and they're more reliable than most people realize.
Running 60+ days continuously
Meta's algorithm kills underperforming creative within 2–4 weeks. An ad still active after 60 days is almost certainly profitable for the advertiser. After 90 days, it's a workhorse.
Multiple variants in the same campaign
If the same brand is running 3–10 ads with slightly different copy/visual variants, they're A/B testing with budget. The brand has a working baseline and they're scaling.
5+ active ads from the same brand
A brand maintaining 5+ active campaigns simultaneously has creative testing maturity. They've figured out what works and they're expanding it.
Running across multiple countries
An ad active in US + UK + CA + AU simultaneously is being scaled geographically. You don't scale a loser. Multi-country active = high-performance creative.
Just started (under 14 days)
Could be a winner. Could be a test about to die. No signal yet — come back to it in 30 days. Don't model your creative on something that hasn't proven itself.
Single ad, single country, brand-page destination
Most likely a brand-awareness or boost-post ad. Not a performance ad. Useful for studying aesthetic but not for copying tactics.
Destination is the homepage (not a product page)
Brand-awareness or general traffic campaign. Performance ads almost always go to a specific product page with a price visible. Filter accordingly.
No price or CTA visible in the creative
Brand-awareness signal. Real conversion ads almost always show a price, an offer, or a strong CTA. If those are missing, the brand isn't optimizing for direct response.
The Legal Line: What You Can and Cannot Do
I covered this in depth in the Rival Mode post, but here's the short version specific to Meta Ad Library research:
✓ Fully legal:
- Browsing the Library
- Screenshotting any ad for personal/internal research
- Studying composition, lighting, color, mood, hook structure
- Creating mood boards from screenshots
- Recreating the visual style with your own product
- Drawing inspiration from copy structure (not copying verbatim)
✗ Not legal:
- Republishing their actual image as your ad
- Copying their logo or trademarked elements
- Verbatim copying of their ad copy / tagline
- Using their recognizable model/brand spokesperson
- Implying endorsement or affiliation
The simple test: if your final ad has your product, your brand, your logo, and your copy — and only the abstract visual style is borrowed (composition, palette, lighting feel) — you're on the same side of the line as every art director in the industry. Style transfer is allowed. Image republication is not.
The Tools That Extend Meta Ad Library
Meta Ad Library is excellent but it has gaps. Three free things it does poorly, and the tools that fix each:
1. Historical search (ads that stopped running)
The Library only shows currently active ads (with limited exceptions). If you want to study the evolution of a brand's creative over 12 months, you need a paid archive. Options:
- AdSpy — biggest commercial archive, ~$150/mo, good if you research multiple verticals weekly
- Dropispy — dropshipping-focused, ~$30/mo, surfaces winning dropship creative
- PiPiADS — TikTok-focused but covers Meta, ~$77/mo
- MinEA — newer entrant, EU-focused
For 90% of solo founders: skip these initially. Use the free Library to build a habit. Upgrade only when you've outgrown it.
2. Engagement signals (likes / shares / comments)
Meta hides social engagement on most Library ads. Paid tools estimate it. If you need this signal regularly, AdSpy is the most accurate. Otherwise, treat the ad's longevity as your performance proxy and move on.
3. AI-tagged "top performers by category"
Some paid tools (Foreplay, BigSpy) tag ads with AI-generated category and trend labels. Useful for trend-spotting. Not essential for the workflow above.
The Handoff: From Found Ad to Your Ad
Finding a winning reference is step one. Step two is turning it into your version. There are three ways to do this:
Hire an agency
Send the reference to an agency. Wait 2–3 weeks. Get an interpretation that costs $2,500–$6,000. Hope the brief got translated right.
Brief a freelance designer
$300–$800, 5–7 days. Depends entirely on whether the designer can read visual recipes accurately. Most can't reproduce premium lighting from a reference.
Run it through Rival Mode
Upload the reference + your product photo to AdLoft's Rival Mode. 8 seconds. ~$0.60. Recipe transferred deterministically.
Most of the customers I talk to do all three at different points. Agency for the hero brand campaign once a quarter. Freelancer for the occasional bespoke piece. AdLoft for the daily creative testing cycle where speed and cost compound into a real workflow advantage.
The Weekly Habit That Compounds
If you take one thing from this post: build a 30-minute weekly research block.
Every Monday morning, before you check Ads Manager:
- Open Meta Ad Library
- Search your top 3 category keywords
- Save 5 new references that look strong
- Add them to a Google Drive folder labeled by month
- Quick-scan the references you saved 30 days ago — which are still active? Those are the proven winners
- Pick one to recreate this week with your product
After 90 days you'll have a curated library of 60+ proven references in your exact category. That's a creative asset more valuable than most brands' entire art direction history. And you built it for free.
Common Mistakes to Avoid
Mistake 1: Only researching the household-name brands
Glossier, Aesop, and Bombas have famous aesthetics — but they're also already imitated to death. The brands worth studying are the mid-size ones running aggressive performance creative right now, with budget but without name recognition. Search by category, not by brand name, and you'll find them.
Mistake 2: Treating brand-awareness ads as performance references
Some of the most beautiful ads in the Library are brand campaigns that aren't optimizing for clicks. They look amazing. They don't convert. Use the destination URL check — performance ads go to product pages with prices. Brand ads go to homepages or vanity landing pages.
Mistake 3: Cloning aspirations you can't deliver on
If your product is a budget item and you clone a luxury ad style, the visual-product mismatch will hurt you. Buyers can smell it. Match your reference tier to your actual product tier — or only go one rung up. Don't try to look like Hermès when you sell at Zara prices.
Mistake 4: Researching in isolation from your own creative testing
The Library is most valuable when you're actively running ads and comparing notes. Pure research without execution feedback gets you a hobby, not a creative strategy. Make sure every reference you save eventually becomes a test in your own Ads Manager.
Mistake 5: Forgetting to study the ad copy alongside the visual
Most posts on this topic focus on the image. The copy hook is half the story. The Library shows the headline, primary text, and CTA for every ad — study the structure (not the words) and you'll get free A/B-tested copywriting templates.
Categories Where This Pays Off Fastest
Certain product categories have outsized payoff from Meta Ad Library research because their visual conventions are codified and the active reference pool is deep:
- Skincare: Saturated with sophisticated performance creative from $50M+ brands. Best learning category.
- Perfume / fragrance: Editorial luxury conventions, easy to reference, hard to brief — perfect Rival Mode territory.
- Cosmetics / makeup: Massive testing budget concentrated in this vertical. Fresh creative weekly.
- Food & beverage: Mood/lighting is everything; references unlock what prompts cannot.
- Handbags & accessories: Premium DTC aesthetic is intensely visual-driven.
- Jewelry: Macro + ambient is impossible to brief in text; reference transfer wins here.
- Facebook Ads and Instagram Ads generally: The Library IS your research engine for these platforms by definition.
Niches with thin Library coverage (fewer winning references): industrial B2B, regional restaurants, professional services. Less valuable for this workflow.
From Research to Shipped Creative — The Full Loop
Putting it together. The complete weekly loop:
- Monday morning: 30 min in Meta Ad Library. Save 5 references.
- Monday afternoon: Pick the strongest reference. Open AdLoft's Rival Mode. Upload reference + your product. Generate.
- Tuesday: If needed, run 2–3 more generations to lock the variant you like. Total cost: ~$2 in credits.
- Wednesday: Launch as a new ad in Ads Manager. Budget $20/day.
- Next Monday: Check performance. If it's beating your baseline, scale. If not, save the reference's recipe lessons and try a different reference. Repeat.
Done weekly, this loop replaces what most brands pay agencies $2k+/month to do badly. It runs on free research + ~$10–$15/month in AdLoft credits + ~$80/week in ad test budget.
If you don't have the product photos to feed this loop yet, run your source through our free Photo Grader first to make sure your input is good enough. Garbage source images can't be saved by clever references — even the best AI product photography tools need a usable starting point.
Frequently Asked Questions
Is it legal to use Meta Ad Library for competitive research?
Yes — Meta Ad Library is a public transparency tool launched in 2019 specifically to make political and commercial ads accessible to anyone. Browsing, screenshotting, and analyzing ads is fully permitted. What you do with the screenshots matters legally: studying composition, lighting, and visual strategy is fair use research. Republishing the actual image, copying trademarked logos, or reusing copyrighted ad copy is not.
How do I know if a Meta ad is actually winning?
Three strong signals: (1) The ad has been running 60+ days continuously — Meta kills underperforming creative fast, so longevity is the cleanest proxy for performance. (2) The brand is running multiple ad variants concurrently — signals they have testing budget and a working baseline. (3) The same brand has 5+ active ads across the library — signals creative testing maturity. Cross-reference these to filter out vanity-only and brand-awareness ads from real performance ads.
Can I see how much a competitor is spending on Meta ads?
Only for political and social-issue ads — Meta is required to disclose spend ranges for those. For commercial DTC ads, spend is not publicly visible. You infer spend through proxy signals: ad longevity, variant count, geographic spread across countries, and whether the brand maintains presence across multiple ad sets simultaneously.
What's the difference between Meta Ad Library and paid ad spy tools?
Meta Ad Library is free, official, and complete — every active ad in the platform is visible. Paid tools like AdSpy, Dropispy, and PiPiADS add historical archives (ads that stopped running), engagement signals (estimated likes/shares), advanced filters (by funnel stage, by landing page tech), and AI-generated tagging (top-performing creatives by category). For 90% of solo founders, Meta Ad Library alone is enough. Paid tools become worth it once you're researching multiple verticals weekly.
How often should I research competitor ads?
Weekly for active campaign managers. Monthly for solo founders. Quarterly for brand strategy reviews. The most valuable habit is keeping a folder of 20–30 reference ads in your category, updated continuously — so you always have fresh inspiration when running new creative tests.
Why are most of my Meta Ad Library searches returning low-quality ads?
Three common causes: (1) Your keyword is too generic — narrow it (e.g. "organic skincare for dry skin" vs "skincare"). (2) Country filter is set to a low-AOV market — try US/UK/CA/AU for premium brands. (3) You're looking at brand-awareness ads instead of conversion ads — filter by "see ad details" and check the destination URL. Ads going to product pages with prices are conversion ads. Ads going to homepages or content articles are brand ads.
10 free credits on signup · No credit card · 8 seconds per output