How to plan a seasonal campaign from last year's ads with Adspy
This Adspy playbook is for teams preparing Black Friday, holiday or summer campaigns. It returns last season's ads on an occasion like Black Friday, grouped by offer and hook, with the date each brand started. Everything it reads is included on every plan, within fair use.
- Cost
- IncludedEverything it reads is included on every plan, within fair use
- Where
- Claude, Claude Code, Codex or any remote MCP clientThrough Adspy's MCP server
- Sources
- Meta and Google adsPublic ad libraries, YouTube ads included; Starter reads its own source's ads, and the TikTok source has none
- Output
- Last season's ads by offer and hookAnd a shared collection of references
Show me the Black Friday ads brands first ran from November 1 last year, oldest first. Group them by offer and hook, tell me which brands started first, and save ten to discuss to a collection called "Black Friday ideas".
What does Adspy do, step by step?
Your AI runs 3 steps with Adspy's tools. Each call is listed with its arguments and the most it can cost on the published tariff; free and included calls use no credits.
Find last season's ads
Your AI searches the ads that name the occasion in their copy, transcript or AI reading, first seen from the date you give, oldest first.
search_creatives{ query: "black friday", since: "2025-11-01", sort: "oldest", limit: 100 }Included
Read their offers and hooks
It reads each ad's AI tags and analysis: the offer, the hook and the format.
get_creative_analysis{ hashes }Included
Keep the references
It saves the ads you pick to a collection your workspace shares.
save_creatives{ name: "Black Friday ideas", hashes }Free
This exampleIncludedeverything it reads is included on every plan, within fair use
What comes back?
Adspy's answer arrives in the chat, with the ads, posts and figures it rests on, so you can open each one.
- Last season's ads on the occasion, oldest first, with their dates
- The offers and hooks that recur, with counts
- When each brand started, from the dates Adspy first recorded
- A shared collection of the references you keep
No sample is shown here: the library snapshot behind the other playbooks' samples cannot show this answer, and a made-up one would mislead.
Good to know before you run it
- The search matches the words of the occasion, so an ad that never names it is missed: try the words brands use, like "cyber week".
- A date is when Adspy first recorded an ad, not its exact start, and an offer that recurs is not proof it worked.
- Starter reads its own source's ads only; the TikTok source has posts, no ads.
Questions about this playbook
Can I keep it to my category?
Yes. Name the brands, or add a category word to the prompt: your AI reads only those ads.
Is it included in my plan?
Yes. Reading ads, their AI analyses and saving them are included on every plan, within fair use.
Have a brand in mind?
Add Adspy to your assistant with this URL, sign in, and ask about its ads. You need no ad account.