Network science
Social graph
A social graph is the network of people and the relationships between them, modeled as nodes and edges. The term went mainstream with Facebook around 2007.
A social graph is a map of who knows whom: every person is a node, every relationship an edge. The mathematical machinery comes from graph theory and decades of social network analysis, but the phrase itself broke into everyday language around 2007, when Facebook began describing its product as 'the social graph' — and when programmer Brad Fitzpatrick published his influential essay 'Thoughts on the Social Graph', arguing that this map of human relationships was too important to be owned by any single company.
Once relationships are modeled as a graph, structure becomes visible that no contact list can show. Clusters reveal your separate social worlds — colleagues, family, the climbing gym. Bridges reveal which single relationships connect those worlds. Distance reveals who you could reach through a warm introduction rather than a cold one.
The unresolved question Fitzpatrick raised is ownership. The big platforms each hold a partial, proprietary copy of your social graph and monetize it; you cannot export the structure, only fragments. A personal, self-owned social graph — your people, your notes, your history, stored where you control it — is the counter-model, and it is the one a local-first personal CRM is built around.
From sociograms to 'the social graph'
Mapping relationships as diagrams predates the internet by decades: Jacob Moreno drew the first sociograms of school classes in the 1930s, and by the 1990s social network analysis had a standard toolkit — Wasserman and Faust's 1994 textbook 'Social Network Analysis: Methods and Applications' remains the field's reference. What changed in the 2000s was scale and ownership. Social platforms turned the abstract graph into live infrastructure: friend lists became edges in a database, and recommendation engines, news feeds and ad targeting all became graph algorithms. Brad Fitzpatrick's August 2007 essay marked the moment developers started talking about 'the social graph' as a layer of the internet itself. He catalogued the problem precisely: every new app forced users to re-find their friends from scratch, because each network hoarded its own copy of the graph, and he argued for a decentralized, portable graph that users could carry between services. Nearly two decades later, that portability still doesn't exist — which is why the term now carries both meanings: a neutral data structure, and a contested asset.
What your personal social graph can tell you
You don't need a platform's data to think in graph terms about your own life. Sketch your forty most important people and draw lines between those who know each other, and three structural facts jump out. First, clustering: your network is almost certainly a handful of dense islands (work, family, old friends, a hobby) rather than one even web. Second, bridges: a few specific relationships are the only paths between islands — if one lapses, two whole worlds lose contact with each other. Third, redundancy and reach: within a cluster, news travels without you, while between clusters it travels only through you. These observations have direct consequences — bridge relationships deserve disproportionate care, an entirely single-cluster network is fragile to job loss or relocation, and the person 'two hops away' is usually reachable through one warm introduction. Network scientists call these patterns density, brokerage and path length; you can use them without ever computing a number.
Owning your graph: the local-first alternative
The social graphs that exist today are platform-shaped: LinkedIn knows your professional edges, Instagram your social ones, WhatsApp your family — and none of them will hand you the full structure, because the graph is the product. Fitzpatrick's vision of a user-owned, portable graph found a partial answer not in a protocol but in a category of software: tools where the canonical copy of your relationship data lives with you. Endearist takes this position literally — your contacts, interaction history, notes and relationship structure are stored local-first on your own device, exportable in open formats, and never used to train recommendations or sell reach. The trade-off is honest: you do the curating that platforms automate. In return, the map of the most personal data you have — who matters to you and why — belongs to you rather than to an advertising business.
Frequently asked questions
- Who owns your social graph?
- Legally, each platform owns its database of your connections; you typically have a right to export your own data (e.g. under GDPR), but exports rarely include the relationship structure in reusable form. Practically, your graph is fragmented across services, each holding a partial copy. The only complete, portable version is one you maintain yourself — in an address book, a personal CRM, or plain files you control.
- What is the difference between a social graph and a contact list?
- A contact list stores nodes only: names, numbers, emails. A social graph also stores edges — who is connected to whom, and what kind of relationship each edge represents. The edges are where the insight lives: clusters, bridges, introduction paths and isolated contacts are all invisible in a flat list and obvious in a graph.
- Is 'social graph' the same as 'social network'?
- They overlap but aren't identical. 'Social network' names the real-world web of relationships (and, colloquially, the platforms built on it). 'Social graph' names the data model: the formal representation of that web as nodes and edges that software can store and query. Researchers analyzed social networks for decades before anyone called the underlying structure a social graph.
Last updated: 2026-06-10
Tend relationships, not records.
Endearist is a local-first personal CRM. Free up to 25 contacts.
Start free