More than half of new launches ask you to act and give you no evidence at all
70% of the 1,086 launches measured make a conversion ask. 55% of those give you no reason to say yes.
Measured studies of launch and funnel UX.
12 measured references that move with the corpus, and 3 frozen studies. Every number is deterministic, no model in the loop.
Studies
Dated snapshots. A cohort measured once and deliberately not moved, so the figures in each stay quotable against the day they were taken.
Score only the landing page and you get 64% of funnels wrong
Score only the entry page and the detected-tactic count is wrong on 64% of funnels.
We X-rayed 94 recent launches. 74% ask you to trust them with empty hands.
The median launch page is nearly empty, and almost nobody makes an argument for itself.
Measured references, one per section type
Each states what the corpus holds for one kind of section — the typical value, the middle half of the range, and how common each element is, with the denominator on every number. They recompute on every build, so a figure quoted from one carries the date it was read.
Measure your own