What public data actually reveals about digital trust

14 July 2026Research8 min

We looked at a year of resolved entities to find out which public signals predicted a bad outcome, and which ones people believe in without evidence.

01
The signals that carried weight
Across a year of resolved entities, the strongest predictors were boring: how long a domain had existed, whether registry records agreed with each other, and whether contact details appeared in more than one independent source. None of them are clever. All of them are checkable.
02
The signals that did not
Several inputs teams routinely weight heavily contributed almost nothing once the boring ones were accounted for. Social presence was the clearest example: its apparent predictive power collapsed the moment domain age was controlled for.
03
Why agreement beats volume
Two sources that independently agree moved our confidence more than ten sources that all derive from the same upstream record. Counting sources is easy and misleading; establishing independence is harder and is what actually matters.
04
What we changed
We reweighted toward corroboration and exposed source independence in the response, so a customer can see whether a high confidence number came from genuine agreement or from an echo.

The signals that carried weight