Felt Important vs Actually Predictive
What you swore mattered against what actually closed deals.
the learner's gut-ranked weights and the backtest's statistically derived weights plot on the same bar chart — the criterion everyone 'just knows' matters often scores lowest against real outcomes
What goes in, what comes out
Learner ranks 10 firmographic and behavioral criteria by how predictive they believe each is of a deal closing, building a weighted ICP score. The engine backtests that exact weighting against a synthetic closed-won/closed-lost dataset the learner never saw.
rank and weight the 10 criteria; submit the ICP scoring model
a backtest accuracy score against real closed-won data plus a ranked list of which criteria were actually predictive versus which the learner over-weighted on instinct
criteria set (10 seeded, e.g. company size, industry, trial-usage depth, inbound source); dataset size (200/1000 synthetic accounts); noise level
Module 'Positioning Backed by Retrieval' L1 lab + B2B portal 'prove your ICP' onboarding tool
Concepts click when you open the machinery.
Three labs are already live and free — the same hands-on style this playground brings to its module.