See what is selling to shoppers right now — and the exact words doing the selling.
Carry the brief out of the browser.
Keep it, and bring it back later.
No saved briefs yet.
Tune the run.
Every format, not just the three.
Send it straight off the screen.
Your name on every export.
Up to ten in one go.
So the pages don't all read the same.
Versions 1 and 2 write the brief on our account, so they're capped by the month. Paste an API provider key of your own and we stop counting. The key is kept in this browser only, and is never sent anywhere to fetch a page.
No key saved. The reading and writing runs on our account.
Skip the clever thing and the new thing. Look only for what is working in the market right now — at companies that spend more testing one headline than this whole project costs.
If a finding makes you say “everyone does that” — that is the finding. Obvious and everywhere is what you came for.
Vague answers give vague research. Write these five lines before any searching starts.
The adjacent aisle line does the most work. A roofing company and a mattress company are both considered purchase, anxious buyer, one big decision, warranty matters — so mattress pages are a fair source of patterns for the roofer.
Start with the current Fortune 500. Delete every company that does not sell to a shopper. Write down the counts as you cut — the counts are what make the list defensible.
Huge testing budgets. What they do is where the market sits.
The majors move last. These brands show what is arriving next.
A pattern that appears in both tiers is not a trend — it is the current standard, and the site should almost certainly adopt it.
Take five pages from each brand: homepage, top category page, one product page, the checkout entry, one live campaign. Fetch them live — training data is older than the market. Every capture carries a URL and a date, and anything you cannot date inside twelve months is labelled history, not heat.
Word-for-word text stays in the swipe file as evidence. It never goes on your site.
A pattern is something you can put a number on: how many Tier A brands, how many Tier B, three examples with URLs, roughly how old it is, why it works on shoppers, and whether this aisle supports it.
Minimum haul: 12 word patterns, 6 offer patterns, 6 proof patterns, 6 communication-style patterns. Fewer means not enough pages were read.
Plus a live word list — the words this aisle uses this year for cheap, fast, guaranteed, easy, new, popular, limited, and the customer. Words age out. The report says what replaced what.
Everyone skips this part, and it is the whole point. Popular is not proof. Selling is proof. Every pattern you recommend must first be tied to money, or to years of surviving on a live page.
The survivorship warning stays in view: these companies win for many reasons, not just this one pattern. A pattern is a possible cause, never a proven one — and no lift number gets written that wasn't measured.
All three prices are struck through for the beta. Join now and you stay free after it ends. Same app, same one job — the higher versions just let the brief travel further and run on your own compute.