1. Classify/
  2. Directory/
  3. Classification/
  4. Jev labels 724 StealAds from 37 brands, 8,724 judgments

Jev labels 724 StealAds from 37 brands, 8,724 judgments

Author: Matthew Berman

Classification

Matthew Berman

Classify this use case

Classification
Cost
9 cents of tokens
Latency
216 ms median per ad, 40 seconds for 724 ads
Result
8,724 judgments across 724 live ads from 37 brands

Ask Jev to label hook, format, offer, CTA, awareness stage, and landing-page mismatch on each of 724 live ads from 37 brands.

StealAds sent each live ad to Jev as a teardown. Jev labelled the hook, format, offer, CTA, and awareness stage, and whether the landing page mismatched the ad.

The published pass covered 724 ads from 37 brands. It finished in 40 seconds, used 9 cents of tokens, and wrote 8,724 judgments at 216 ms median per ad.

The same breakdown will be available in StealAds and an MCP.

How it works

  1. Collect live ads from 37 brands
  2. Ask Jev for hook, format, offer, CTA, awareness stage, and landing-page mismatch on each ad
  3. Roll the labels into one category teardown
  4. Ship the same pass in StealAds and an MCP