Instagram Trend Finder
Finds the reels genuinely outperforming their own account, then turns the winners into briefs you can film.
The problem
Everyone says "study what is working in your niche." In practice that means scrolling for an hour, saving eleven reels, and being no clearer than when you started.
The deeper problem is that raw view counts lie. A million-view reel from a two-million-follower account is a normal Tuesday. A 250k reel from an account that usually does 8k is the actual signal, and it is invisible if you sort by views.
What I built
A pipeline that pulls recent reels across a set of accounts and scores each one against its own account median rather than against the whole platform. A reel doing 611× its account's normal performance surfaces above one with more absolute views.
The score blends an outlier weight with a velocity weight, so something climbing fast ranks alongside something that already peaked. The dashboard is a triage surface: sort, select the ones worth stealing the structure of, export the selection.
Selected reels become Markdown briefs — hook, structure, why it worked — so the output is something you can film from rather than a folder of saved links.
See it working
This one runs on private data, so there is no public demo.
What changed
Normalising against each account's own median was the whole thing. It turns "this got a lot of views" into "this beat what this account normally does," which is the only version of the question that tells you anything.
- Top outlier found
- 611×
- Reels triaged
- 99
- Not a bookmark
- Brief
Next project
Upwork Screening Pipeline
Scrapes the job board, throws away the ones that were never a fit, and drafts proposals only for the rest.
The easiest next step
Want something like this for your workflow?
Tell me what you are doing by hand right now. That is genuinely enough to start from.