Six Degrees of Separation and Why Entity Graphs Get Noisy Fast
You put three people on a graph. You expand one of them. Suddenly there are 60 nodes and 300 edges.
You didn’t do anything wrong. The graph isn’t broken. You just ran headfirst into one of the most well-known properties of interconnected data — and it has a name.
Six Degrees of Separation
In 1967, psychologist Stanley Milgram ran an experiment. He asked random people in Nebraska to get a letter to a specific person in Boston — but they could only mail it to someone they knew personally, who would then forward it to someone they knew, and so on.
The average chain length? About six.
The idea stuck: any two people on Earth are connected through roughly six intermediary relationships. It’s been tested repeatedly since then, including a 2011 study by Facebook that found the average distance between any two of its billion-plus users was 4.74.
This isn’t trivia. It has direct consequences for anyone using entity graphs — whether you’re running a financial investigation, conducting due diligence, tracing digital assets, or reviewing a document production for litigation.
What This Means in Practice
When you expand an entity on a graph, you’re walking one of those chains. A Person connects to an email address. That email connects to another Person. That Person connects to an Organization. The Organization shares a phone number with a third Person.
Each expansion — each hop — brings in more entities. And because the chains between any two people are short, you reach a large portion of the network quickly.
This is the fundamental difference between entity graphs and other types of graph tools.
The Contrast: Cryptocurrency Graphs
If you’ve used cryptocurrency attribution tools like Chainalysis Reactor or TRM Labs, you’ve experienced a different kind of graph. Crypto graphs have structure. A wallet sends to another wallet through a transaction. There are clear walls — the data can only flow through defined channels. The graph grows, but it grows in predictable ways.
Entity graphs don’t have the same walls.
A Person can connect to emails, phone numbers, usernames, addresses, organizations, crypto wallets, passport numbers, case numbers — and each of those can connect to other people and organizations through shared identifiers. The data has no inherent boundaries. A location like “Miami, FL” might appear in hundreds of documents across dozens of matters, connecting people who have no actual relationship. Nonetheless, locations are of critical importance to explorations, matters, and investigations. Alone, their meaning is less clear and they are less helpful. But combined with other identifiers, they can make all the difference.
The result: entity graphs can become visually overwhelming within a few expansions. This is not because the data is wrong, but because the data is interconnected by nature.
Managing Complexity Without Losing Signal
Recognizing this complex truth is step one. Building tools to handle it is step two.
There are a few approaches that work, and we’ve implemented all of them in Ingestigate’s Graph Explorer:
Automatic Clustering
When a Person or Organization node has multiple connected identifiers — three email addresses, two phone numbers, a passport number — those identifiers can be folded into a single clustered node. Instead of seeing 7 separate nodes radiating from a Person, you see one compact cluster labeled with key identifiers (which can be changed and customized by the user with a simple right-click of a mouse).
The information is still there. You can expand the cluster to see everything. But the default view keeps the graph readable.
Smart Type Filtering
Not all entity types carry equal weight at every stage of an analysis. Dates, locations, and IP Addresses create connections that can create a lot of extra noise at the graph level — a city name appears in every document from that region, connecting people who have no meaningful relationship.
The Graph Explorer automatically simplifies the view by hiding noisy entity types on expansion, while keeping the most important types visible: People, Organizations, Email Addresses, and Usernames. These are the entities that function as identities — the anchors of any exploration.
Everything hidden is one click away. The Node Types panel shows exactly what’s been filtered and lets you show or hide any category instantly.
User-Controlled Grouping
When a specific type matters in your work — say you’re tracing 50 crypto addresses in a digital asset investigation, or reviewing 30 organizations in a corporate compliance matter — you can right-click any node and group all nodes of that type into a single collapsible group. One click, one node. Expand it when you need the detail, collapse it when you don’t.
The Right Mental Model
The key insight for anyone working with entity graphs is this:
Graph density is a property of the data, not a flaw in the tool.
When you expand a well-connected entity and the graph gets busy, that’s the graph doing its job — showing you that this entity has a large network of connections. The question isn’t “why is this so messy?” but “which of these connections matter to my work?”
And the graph page is more than a read-only view. Right-click on the empty space of the canvas to add an Annotation — a note you can attach to specific entities by drawing connections to them. Remove a node you don’t need, or remove all nodes of a certain type, with a simple right-click. Save the graph to a board and load it on demand, anytime. What starts as an exploratory tool becomes a persistent, visual documentation surface for your matter or case.
The tools exist to help you focus: cluster the identifiers, filter the noise, group what you don’t need right now, annotate what you’ve found. Start with the People and Organizations. Follow the emails and usernames. Expand into other types when you have a specific reason to.
Entity graphs reward focused exploration more than broad expansion. Each hop should be intentional. And when the graph gets dense, that’s your signal to filter, cluster, and focus — not to start over.
Ingestigate’s Graph Explorer is built for work where relationships matter more than individual documents — from financial investigations and compliance reviews to litigation support and digital asset tracing. Start your free trial and see how entity graphs can surface the connections that matter.