Meta Ads Ad Name N-Gram Analyzer
Your naming convention is a goldmine. Paste your ad-level export and see spend, CPA and ROAS aggregated by every hook, creator, format and angle in your ad names, so the next brief is built on what actually wins. Nothing is uploaded; it all runs locally.
The Meta equivalent of search terms mining
On Google you mine search terms to find the words that waste money. Meta gives you no such report, and individual ads rarely have enough conversions to judge on their own. But the ingredients repeat: the same creator appears in eight ads, the same hook in five, the same offer in twelve. Aggregating performance by those recurring name tokens turns thin per-ad data into statistically useful per-ingredient data. It's how you learn that a creator wins regardless of hook, or that an offer only works in UGC, before you spend another month's production budget guessing. Paid tools charge for exactly this analysis; your ad names already contain it.
How to use this tool
- In Ads Manager, set a 30 to 90 day range, view at ad level, and export the table as CSV. No breakdown needed, just totals per ad with spend, results and value.
- Upload or paste it above. The delimiter is auto-detected from your names, or pick it manually.
- Read the table: winners deserve more ads built on that ingredient, dragging tokens deserve scrutiny, and wasting tokens are burning spend with nothing back.
- Download the token report and take it into your next creative brief.
FAQ
No. The export is parsed and analysed entirely in your browser, it never leaves your machine and we never see it. No OAuth, no sign-up.
Meta has no search terms report, but a disciplined naming convention encodes hook, creator, format, offer and angle in every ad name. Tokenise those names against spend and results and you get the Meta equivalent of n-gram analysis: which ingredients repeat across your winners.
The tool still works on whatever structure exists, spaces included, but the cleaner the convention the sharper the read. If this analysis convinces you to adopt one, a simple Format | Creator | Hook | Product pattern pays for itself within a month.
Each token's CPA is compared with the account average from your paste. At least 20% better is a winner, 25% or more worse is dragging. Tokens with spend but zero results at meaningful cost get flagged as wasting.