OUR APPROACH · UPDATED OCTOBER 25, 2026

Find what’s winning.
Work out why. Make yours.

Advertisers keep paying for native ads that make them money and switch off the ones that don’t. Most native ads are gone within days. The ones still running after weeks are earning their budget. That is where we start.

1. How we find what’s winning

Every week we read the most-bought native ads on Taboola, Microsoft, Outbrain, Revcontent and MGID from public native ad libraries. An ad counts as a winner when it keeps running for weeks and is bought on many sites. We sort every ad into a space, like pain relief or Medicare, and group the long-running winners that several different advertisers share.

The numbers behind it

• What we read: up to 2,000 of the most-bought US native ads each week, across Taboola, Microsoft, Outbrain, Revcontent, MGID, Yahoo and smaller networks, from public native ad libraries. Weekly records start with the week of October 5, 2026.
• What counts as winning: an ad still running after 30 days, or bought on more sites than most of the ads around it. Nobody keeps paying for an ad that loses money.
• What counts as a formula: a pattern at least 3 ads from at least 2 different advertisers share. One advertiser’s habit is not a formula.
• When a space gets a page: at least 40 ads from 15 advertisers that week, with 3 or more formulas (20 ads from 8 advertisers for a narrower topic). Below that, there is no new page, and the latest full week is shown and labelled as such.
• What we measure: how many days each ad has run, how many sites it was bought on, how many advertisers share a formula, and how long each formula keeps winning week after week. Past weeks stay published, so you can see what changed.

2. How we work out why they win

A winning ad wins for reasons you can name. We break each group down into its winning formula:

• The hook: what stops the reader mid-scroll.
• The emotion: what the reader feels in that second.
• The persuasion: the principle doing the work, such as social proof, authority, liking, scarcity, reciprocity, commitment or curiosity.
• The mood: the feeling of the whole picture.
• The kind of photo: what is in the frame and how it was taken.

An example. A night-time flash snapshot of friends laughing round a table. The hook: it looks like a photo from a friend’s phone, not an ad. The emotion: belonging. The persuasion: social proof and liking, people like me having a good time. The mood: warm and unposed. The kind of photo: a candid flash shot, a little messy. Change the product and the people, keep those parts, and the ad still works.

Reading ads this way comes from years of running native ads and thousands of ad tests.

3. How we make new ads from it

On a plan, every Monday we make brand-new ads for your business from the winning formulas in your space: new pictures and headlines for your product, finished in every native size. Same psychology, new pictures. We never copy anyone’s ad or use another advertiser’s pictures, names or claims.

Every ad is checked against the networks’ rules, and the winners already running set the bar: what they run is fair game in the same form, never more extreme. Nudity, gore, real recognisable people and official seals are always out.

When the evidence is thin

Some spaces have only a few advertisers in a given week. We say so plainly and never dress thin evidence up as a sure thing.

About the examples

The examples on our homepage use fictional companies and AI-created images so no real advertiser is shown. What you see for your own business uses live ads from public native ad libraries, with the dates each ad was seen.