The short answer
Everything you need is public: Meta's Ad Library shows every active ad, its start date and its versions, and the landing page is one click away. Record the hook, angle, format, offer and page for each ad. Then borrow structure, not words: of 1,680 winning video openers we checked, only 8 are shared word for word between different companies.
Everything you need is already public
"Spying" sounds shady, but nothing here requires a trick. Meta built the Ad Library so that anyone can see the ads a Page is running. Every other piece of the funnel is public too.
| What you can see | Where | What it tells you |
|---|---|---|
| Every active ad and its creative | Meta Ad Library | The current message mix |
| Start date and versions | Ad Library card | Which ads survived and which are being scaled |
| Comments and reactions | The ad post itself | Objections, desires, the words buyers use |
| The landing page | Click the CTA | Offer, price, page type, proof |
| The checkout and upsells | Go one step further | AOV strategy |
| Emails and retargeting | Sign up, visit, leave | Follow-up sequence and remarketing creative |
What you can't see: spend, results and targeting outside the EU. Everything else is a click away, which is why the work is in organizing it, not finding it.
Who to watch (and how many)
Not every competitor deserves the same attention. On the Ad Radar board (2,443 active Meta ads with $50k+ in estimated spend, October 2026), winning ads are concentrated in a few advertisers:
749
brands with at least one $50k+ ad
396
of them have exactly one
47
have ten or more
Most brands find one winner and ride it. A small group runs ten, twenty or sixty at once. Those high-volume advertisers are where you learn the most per hour, because you can compare their ads against each other and see what they keep.
Build a watchlist of five to ten names:
- Two or three direct competitors that sell what you sell to the same buyer.
- Two adjacent brands that sell something different to the same buyer. Their angles transfer and nobody in your niche is copying them.
- One or two volume leaders in your category, the ones with dozens of active ads.
What to record for every ad
A screenshot is not a record. When you find an ad worth keeping, log these fields. It takes two minutes per ad and makes the file searchable months later.
| Field | Example | Why |
|---|---|---|
| Brand and Page | Chef Preserve, own Page | Partner Pages behave differently |
| Start date and days live | 371 days | Survival |
| Versions | 2 | Scaling |
| Format and length | UGC selfie video, 53s | Production cost and style |
| Hook (first sentence, verbatim) | "Here is my secret on how I..." | The part you will test |
| Angle | Accidental discovery | The argument behind the hook |
| Awareness level | Solution-aware | Who it is talking to |
| Offer | 5 times longer, 5-second seal | What makes them click |
| Landing page type | Dedicated landing page | Where the click lands |
| Your one-line read | Bulk shopping habit, not "vacuum sealer" | The insight, in your words |
Here is that ad on the board:
Here is my secret on how I shop in bulk and preserve our food for weeks.
- Est. spend
- $1.9M
- Days live
- 371
- Format
- Video, 53s
- Variants
- 2
The product never opens the ad. A shopping habit does. That is the line worth recording, because it shows how to sell an appliance without leading with the appliance.
The ten fields map closely to the ten dimensions Ad Radar tags on every ad (hook pattern, hook mechanism, angle, awareness, proof, production format, video type, landing type, niche, CTA). If you build your own sheet, steal the structure from our swipe file guide.
A weekly routine that fits in one sitting
Pick a fixed day. For each brand on the watchlist:
- Open their Page in the Ad Library, active ads only, your country.
- Note any ad that crossed 30 days since last week. That is a test that survived.
- Note any ad whose version count went up. That is a winner being scaled.
- Click through to the landing page of anything new and screenshot the offer.
- Read the top 20 comments on the oldest ad. Copy the objections word for word.
- Log everything in the same sheet, one row per ad.
The full checklist with a template is in competitor ad research: a 30-minute weekly process. To automate the noticing part, see automating competitor ad monitoring.
How to avoid copying (and why winners don't)
Here is something we didn't expect. We took the first spoken sentence of every video ad on the board that has a transcript, 1,680 openers, and looked for identical lines across different companies.
Only 8 openers appear word for word at two or more unrelated companies. That is 0.5%. And the ones that repeat are mostly generic: urgency lines like "this is going to be the last day to get it for crazy prices like this", a trending audio clip, or story openers reused across affiliate funnels.
This is going to be the last day to get it for crazy prices like this.
- Est. spend
- $674K
- Days live
- 78
- Format
- Video, 14s
- Variants
- 4
The same urgency line opens ads for four different companies on our board. It is a format, not an idea, which is why it travels. The body copy carries the real angle.
Everything else that works is original at the word level. What repeats is the structure: a confession opener, a specific number, a callout to one kind of buyer, a root-cause explanation. That is what you should take.
I caught my husband cheating with a girl half my age.
- Est. spend
- $1.5M
- Days live
- 97
- Format
- Video, 481s
An eight-minute story that starts as a personal drama and ends at a product. The same opening line appears in another company's ad. If you borrow the structure (life event, then the hidden cause), write your own event.
A simple test before you brief anything inspired by a competitor:
- Would a customer confuse the two ads? If yes, you copied.
- Could you explain the idea without quoting them? If not, you copied the words, not the idea.
- Did you change the specific? Their number, their character, their scene should all become yours.
- Does it fit your product's truth? A borrowed claim you can't back up is a compliance problem, and it is yours, not theirs.
Turn what you watched into briefs
Spying is only useful if it changes what you make next. Once a month, read your log and answer three questions:
- Which hook patterns appear in the ads that survived past 90 days?
- Which angle is everyone in your niche using, and which is nobody using?
- Which landing page type do the long-running ads point to?
Write one brief per answer. Our guide to reverse-engineering a winning ad breaks one ad into a brief in 30 minutes, and how to find winning Facebook ads covers the signals that decide which ads earn a place in the log. If you would rather skip the logging, Ad Radar lets you follow any brand by pasting its Ad Library link and alerts you when a new ad reaches the top tier.
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.