Pivot scores
Pivot ranks what it finds with a few scores. They measure the evidence Pivot observed, not who owns an asset. In the node panel, hover or focus a score's name to see what it measures.
What does confidence mean?
Confidence tells how strongly the evidence ties a node to your seed, from 0 to 100%. Independent kinds of evidence count more than several pieces of the same kind: a domain that shares your analytics ID and your name servers usually ranks above one that shares three of your IP addresses.
What is relevance to the seed?
Relevance to the seed tells how central a node is to your seed across the whole graph. The most central node is at 100% and the others are measured against it. A node reached through several strong links ranks above one that hangs off a single weak link, so relevance is a good guide to where to pivot next.
How is a link scored?
Each connection has a link score. Select a node and read Why this link to see the terms behind it. A term that changes nothing is hidden, ↑ marks a term that raises the score and ↓ one that lowers it.
| Term | What it measures |
|---|---|
| Rarity | How few hosts share the value in the index. A value shared by millions of hosts says little. |
| Signal reliability | How much this kind of evidence usually says about common ownership. A shared analytics ID says more than a shared IP address. |
| Similarity | For favicons, how close the icon looks to your target's. |
| Value context | What Pivot knows about the value itself. An address at a shared hosting provider, or a value other analysts often mute, weighs less. |
| Timing | Whether the observations overlap in time. Evidence last seen long ago weighs less. |
| Cohesion | For a value many domains share, how many of those domains other evidence also ties to your seed. |
| Certificate diversity | For a certificate, whether the names it lists belong together. A certificate bundling unrelated sites weighs less. |
The same terms explain the percentage next to each row of Correlations.
Why are some sources muted automatically?
Some values link thousands of unrelated customers: CDN and website-builder addresses, parking services, and the name servers registrars use for suspended domains. Pivot mutes them so they do not crowd out the evidence that matters. Muted automatically lists them; choose Show to bring one back into the graph.
What are groups and bridges?
Pivot splits the graph into groups: nodes linked more to each other than to the rest of the graph. In the full graph, turn on Color by group to see them. A bridge ties groups together, and removing it would split the graph. A bridge is often where two operations, or two parts of an organization, meet.
How should I use the scores?
Use them to decide what to look at first. They rank leads and do not prove ownership, and correlation levels are read separately. Before you use a connection, open the node, check Why this link and the supporting observations, and make your own assessment.