THE CREATOR’S PLANNING BENCH / 001
Before you post,
change the perspective.
Test the piece. Imagine the viewer. Give your next batch a better starting point.
An independent planning heuristic. Source claims are unverified. No data is collected.
THE PIECE
RecommendScore
Describe your planned post. Every choice stays on this page.
NEUTRAL / NO STATE RULES
The planning readout
A fresh perspective starts with a few details.
Describe your piece, then run the score.
Probabilities are planning estimates, not platform predictions. Lower skip is better; higher finish is better.
THE CONTEXT
Viewer state simulator
Same piece, different moment. All states are manually selected.
Score a piece to see all six deltas against neutral.
State fit checks
- Practical fixes will appear here.
THE SEQUENCE
Batch rescue planner
Queue two to five pieces from the scoring form. Explore a sequence before committing to it.
- Your queue is empty. Start with the piece above.
A PLAN, NOT A PROMISE
Make the first piece count.
When two high-skip pieces meet, the planner brings your strongest finisher forward and checks the sequence again.
A NOTE ON THE MODEL
Inspired by a claim.
Built for your own judgment.
ForYou Lab is inspired by the source reel by ayoships (Ayo Omolewa) ↗. The reel’s claims about TikTok are unverified. This app does not reproduce TikTok’s actual recommendation model.
Every score is a deterministic planning heuristic, not a measured likelihood or a promise of reach. The app collects no data. It uses only values deliberately entered here, keeps them in memory, and clears them on reload. No device signals are read.
How the heuristic works
Skip starts at 0.55; finish at 0.45. Early hooks, captions, muted readiness, loops and pattern interrupts reduce skip and raise finish. Longer duration adds pressure. Probabilities are bounded from 0 to 1. Like, comment and follow combine a baseline with finish and matching call-to-action bonuses. Staytime is duration × (0.2 + 0.7 × finish − 0.1 × skip), capped at duration.
State rules add modality and attention-window adjustments: Lunch Break prioritizes muted readiness and captions; In Bed rewards spoken hooks and longer attention, with a penalty for text-only hooks; Walking favors audio; In Car requires spoken audio; Desk uses medium attention. Fit checks treat a spoken hook as the proxy for audio carrying the message. Factor lists show the three largest additive contributions before bounds apply. Full constants and rules are documented in score.js.