AI Blocks › Read your ad results
Know which ads to keep, cut and make next after a test
Give it your test results and the ads. Get a verdict per ad, a next-round brief and where the next dollar goes.
Find why your ad test lost and plan the next is a free AI workflow for the job "Read your ad results". It guides your AI through several skills in order, passing each step's output to the next.
Free · No account · Made by AI Blocks · Thinking: high
What it does
A round of ads ends and the numbers are muddy. One ad has a good return on a few sales, another spent a lot and did little, and you cannot say whether the idea or the execution failed. Often the round tested one idea ten times, so there was never a real comparison. The first step splits every ad into its message and its look, and writes the one idea behind each. The second counts distinct ideas, not ads, and shows how many were repeats. The third reads the results by ad set and by concept, flags false winners and writes the brief for the next round. The fourth turns that brief into a Meta test layout with a minimum spend. The fifth ranks channels by new-customer revenue per extra dollar. It needs real spend and results per ad. With only a handful of sales in total, the reads are…
You give it
- The ads you ran (images, videos and their copy, or a description of each)
- Results per ad and ad set, with spend, purchases, return, clicks and frequency
- Your target return or cost per sale
- Spend and results by channel (optional, but step 5 needs it)
- Rival or reference ads you want compared (optional)
- Brand brain (optional)
- Your question (optional)
You get
- A breakdown of every ad by message and look, with the one idea behind each and the patterns across winners
- A sameness score out of 100, the number of distinct ideas, and which ads to keep and replace
- A verdict for each ad set and ad, false winners flagged, and a read score out of 100
- A Meta test layout for the next round, with a pack plan, minimum spend, naming and scaling rules
- A ranking of channels by the next $10,000, with a stop point for each and a test that could change the ranking
Use it when
- A round of ads finished and you cannot tell what won, what to cut or what to make next.
- Your best ad has stopped working and you are not sure whether to make more like it or something new.
- You are about to add budget and want to know where an extra dollar earns the most new customers.
Steps
- Successful ad pattern analysis: Share successful ads. Get a breakdown of their messages and visuals, with single-element changes to test.
- Ad sameness checker: Paste your live ads with persona, angle, offer. Get idea groups and a sameness score out of 100.
- Ad creative test results reader: Paste ad results by ad set. Get keep, leave or replace calls, false winners and a score.
- Meta test campaign planner: Give it your products, concepts and spend. Get a test campaign layout, ad pack names and scale rules.
- Marketing budget allocation planner: Give it spend and new-customer returns by channel. Get where the next $10,000 goes and a test plan.
How to use it
- Download the handoff pack. It's a zip with HANDOFF.md and every skill file this workflow uses.
- Give the folder to Claude (in Claude Code, or as files in a Claude Project) and say "Follow HANDOFF.md".
- Attach the ads themselves where you can, as images or video stills, and paste the results export beside them.
- Say your target return or cost per sale, so each verdict is against your goal and not a general one.
- Read the false winners and the sameness count before you decide what to brief next.