AdRoastWhat is an ad teardown, and how to do one
An ad teardown is a breakdown of a real live ad and the landing page behind it, scored on how well the message survives the click. It is not a critique of the creative. It is a test of the handoff, and the handoff is where the money leaks.
Most ad feedback stops at the ad. Someone screenshots a LinkedIn ad, calls the hook weak or the image busy, and moves on. That is a creative critique, and it misses the part that decides whether the ad makes money. A visitor does not convert inside the ad. They convert on the page after it. A teardown follows the click and grades the whole journey: what the ad promised, and whether the page kept that promise. The single most useful number a teardown produces is the gap between those two.
Why a teardown beats a critique
An ad can be genuinely excellent, sharp hook, right buyer, credible proof, and still fail completely. It fails when the click lands on a page that changes the subject: a homepage, a 3,000-word blog post, a feature tour, a founder essay. The visitor arrived with a specific expectation and the page answered a different question. A critique of that ad would give it high marks. A teardown gives it a low score, because the teardown grades what the buyer actually experienced, not what the ad said in isolation. When you only fix ads, you keep polishing the half of the funnel that was already working.
How to do one, in four moves
Name the buyer. Who is this ad for, in one sentence: role, company, and the problem they already know they have. Every score after this is relative to that person, not to your taste.
Score the ad's promise. Read the ad as the buyer. What exactly does it promise, and how specific is it? A promise a buyer can picture ("cut alert noise from 11,000 to 100") scores high. A vague brand line scores low.
Follow the click and score the carry. Open the landing page, first screen only. Does it repeat that promise, in roughly the same words, within the first thing the buyer sees? Score how much of the promise survives the handoff.
Write the gap as three fixes. The distance between the ad's promise and the page's follow-through is the finding. Turn it into the three specific changes that would close it. Not "improve the page". "Put the ad's stat in the H1 and move the demo form below one proof block."
A teardown is not about whether the ad is good. It is about whether the promise survives the click. Most of the time, it does not.Marina Kogan
See real teardowns
The method is easier to trust when you see it run on real, named companies. AdRoast publishes teardowns of live B2B cybersecurity ads and the landing pages behind them, each scored with the exact fixes: read the teardown series. They cover strong ads that die on the wrong page, sharp hooks that hand off to a blog post, and single-product ads that land on a page selling eight.
To get one of your own, AdRoast runs an automated teardown from just your website URL: it pulls your live LinkedIn and Google ads, follows each click, scores the handoff against your buyer, and writes the fixes. The first one is free.
Get a teardown of your own ads
AdRoast pulls your live ads from just your website, follows each click, and scores the handoff against your buyer, with the exact fixes. First roast is free.
Roast your ads free →Keep reading
How to audit your LinkedIn ads against your ICP
A 5-step method to score whether each live ad actually speaks to your buyer.
Clicks but no pipeline: why B2B ads leak after the click
The four places the funnel breaks between the ad and the form, and how to find yours.