What predicts a close in cybersecurity.
Technical evaluation and a crowded field. Competitor mentions cost real points.
The argument
Security buyers self-educate before they will speak to anyone, so third party intent and site behavior carry the most weight of any segment. Documentation and pricing page visits are evaluation, not curiosity. The category is dense enough that a competitor named on a call is materially bad news rather than noise, so the competitor signature is amplified and both competitor signals are made heavier. Cycles are shorter than the regulated segments and evaluations are proof of concept driven, which makes recent activity more predictive and stalls less so. The split moves toward the person side because a security engineer running a bake-off is a real buyer, not an influencer.
The weights
Published, because a score you cannot audit is a score nobody trusts.
66% of the score comes from what the company does, 34% from what named people do. Within each side, these are the shares.
Company signals
Person signals
See it against a real pipeline.
The demo loads with these weights. Move them and watch the deals reorder.