The short answer
A winning product leaves tracks in the Meta Ad Library: ads that stay live for months, multiply into variants and get pushed from many pages at once. On our board of 2,443 winners, ads with 4 to 9 variants have a median estimated spend of $528K, three times the $177K of single-version ads. Food and drink products reach the top tier most often (16%), supplements least (8%).
The five signals that a product is winning
Product research lists tell you what someone thinks is hot. Ads tell you what someone is still paying to sell. We went through the 2,443 live ads on our board (749 brands, all with $50k+ in estimated spend) and looked at which public signals line up with the biggest spend.
$1.12M
median est. spend, ads with 10+ variants
$177K
median est. spend, single-version ads
16
ads under 60 days old already past $1M
12
separate pages pushing one dog food brand
Here is what to look for, in order of how much each one tells you.
1. Variants: the strongest single signal
When an ad works, the advertiser launches more versions of it. New hooks, new cuts, new headlines. Variant count is the closest public proxy for "this made money", and the spend data agrees.
| Variants per ad | Ads | Median est. spend |
|---|---|---|
| 1 | 1,743 | $177K |
| 2 to 3 | 524 | $332K |
| 4 to 9 | 148 | $528K |
| 10 or more | 28 | $1.12M |
Variant count is one of the inputs to our winner score, so the top-tier share climbs with it by design. The spend column is the independent check: it doubles, then doubles again. More on this in ad variants as a scaling signal.
2. Days live
The median ad on the board has been live 120 days. Advertisers rarely keep paying for a loser for four months. A product whose main ad has run 90+ days has passed the test most products fail. The full distribution is in how long winning Facebook ads run.
3. Many pages, one landing domain
This is the signal most product researchers skip. When a brand scales, it starts running ads through creator pages and partner pages, all pointing to the same site. In the Ad Library they look like unrelated advertisers. On our board, 72 landing domains are reached from three or more different page names.
| Landing domain | Page names | Ads on board | Median est. spend |
|---|---|---|---|
| thefarmersdog.com | 12 | 43 | $567K |
| tryrovina.com | 13 | 48 | $497K |
| factor75.com | 11 | 50 | $209K |
| javvycoffee.com | 10 | 17 | $199K |
| dogishuman.com | 6 | 83 | $236K |
| mykitsch.com | 6 | 36 | $193K |
The fresh dog food brand is the clearest case: 12 pages, mostly named after dogs and their owners, 43 ads, and 10 of them in the top tier. That's a product with proven demand and a brand willing to pay creators to sell it.
Get 50% off your first box
- Est. spend
- $3.8M
- Days live
- 50
- Format
- Video, 38s
A dog's page, not the brand's, running a first-box offer for a fresh dog food brand. Past $3.8M in estimated spend in 50 days. If you only searched the brand name, you'd miss it.
One caution. Multi-page domains, taken as a group, don't reach the top tier more often than single-page brands (9% against 11%). Many pages tells you a brand is scaling a product. It doesn't tell you each ad is good.
4. Speed: big spend on a young ad
Most ads take months to build up spend. A few get there in weeks. 16 ads on the board are under 60 days old and already past $1M in estimated spend, and 13 of them reach the top tier. Those are products being pushed hard right now.
I never thought I would like short press-ons at my age, but here we are.
- Est. spend
- $8.4M
- Days live
- 56
- Format
- Video, 59s
- Variants
- 3
Soft gel press-on nails, past $8.4M in estimated spend in under two months, with a buy-one-get-one headline. A product in its scaling window.
5. Ad count per brand
The weakest signal on its own. 125 brands have five or more ads on the board, and 396 have exactly one. A brand with many winners has found a product the market wants, but count also reflects how many creatives a team can produce. Use it to shortlist, then check signals 1 to 4.
Which product types win most often
We tag each ad with the form of the product it sells. Here is how they compare.
| Product form | Ads | Brands | Reach the top tier | Median est. spend |
|---|---|---|---|---|
| Food and drink | 137 | 34 | 16% | $376K |
| Home goods | 86 | 21 | 14% | $247K |
| Gear | 98 | 16 | 13% | $227K |
| Topicals | 172 | 51 | 12% | $214K |
| Apparel | 176 | 27 | 11% | $198K |
| Devices | 219 | 52 | 11% | $234K |
| Supplements | 609 | 143 | 8% | $207K |
The top tier is the top 10% of the board by winner score, among ads live 21+ days. About 10% is the average.
Two things stand out. Food and drink leads on both top-tier share and median spend. Two subscription brands, a prepared meal service and a fresh dog food, account for 93 of the 137 ads. A winning subscription ad keeps paying through repeat orders, which gives the advertiser room to spend more on it. So read this row as "subscription food wins", and note the sample is concentrated. And supplements are by far the most crowded form, with a top-tier share below the board average. There's plenty of money in supplements. There's also plenty of competition for it.
For a breakdown by category instead of form, see which e-commerce categories win most often.
Do you have an itchy dog who smells like corn chips?
- Est. spend
- $10M
- Days live
- 57
- Format
- Video, 46s
- Variants
- 3
The brand with the most ads on our board (63). A pet supplement that sends traffic to a quiz. A strong product with a sharp, specific symptom in the hook.
Reading the ad to judge the product
Signals tell you a product sells. The ad itself tells you why, and whether you could sell it too.
- Is the hook about the product or about the problem? "Smells like corn chips" is a problem anyone with a dog recognizes. Products with a vivid, visible problem are easier to sell cold.
- Does the ad need a demo? Press-ons, luggage and hair tools sell on a 5-second visual. Products that need explaining need longer ads and more proof.
- Where does the traffic go? A quiz or listicle landing page means the brand had to warm the buyer up. A plain product page means the ad does all the work. See product page vs landing page.
- What's the offer? First-box discounts and buy-one-get-one both signal a product with repeat purchase or high margin.
Man, there's nothing worse than looking clueless at the airport.
- Est. spend
- $4M
- Days live
- 35
- Format
- Video, 44s
- Variants
- 3
A carry-on bag brand with 26 ads on the board. This new version passed $4M in estimated spend in 35 days. The product shows its features in seconds, which is why it scales on video.
A product research routine, step by step
- Pick a category and pull every ad in it with 90+ days live.
- Sort by variant count. Keep anything with 3 or more.
- Group by landing domain, not by page name, so creator and partner pages roll up to the real brand.
- Flag young ads with big spend. Under 60 days and over $500K estimated is a product in its scaling window.
- Read the top three ads per product. Write down the problem in the hook, the proof used and the offer.
- Ask whether you can beat the offer or the angle, not just copy the product. If you can't, move on.
Steps 1 to 4 are filters you can run on the Ad Radar board in a few minutes. For the full process, see how to find winning Facebook ads, and if you're doing this on a small budget, dropshipping ads in 2026 covers the saturation side.
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.