Forget the social graph: pay attention to the work graph

The term ‘social graph’ was popularized at the Facebook F8 conference back in 2007, and helped that company explain what it was trying to do in its core architecture and business models. At first I considered the term as having no real difference — a near synonym — for social networks, with the exception of stressing the ‘rigorous mathematical analysis’ derived from graph theory.

However, over time I came to appreciate that the social graph is actually a larger formulation than social networks: it is a graph (or network) of people as well as social objects — the things that people are talking about, or sharing, that shape the relationships between the people in the social networks.

It turns out that the term was originally offered up by my friend, Jyri Engstrom, the founder of Jaiku, back in 2005, when he wrote Why some social network services work and others don’t — Or: the case for object-centered sociality:

Social network theory fails to recognise such real-world dynamics because its notion of sociality is limited to just people.

Another friend, the cartoonist Hugh MacLoed (@gapingvoid) popularized the term in the years following, as in Social Objects for Beginners.

We’ve seen a few major trends in the consumer side based on the magnetic power of the social graph:

  • photo-centric social tools like Instagram have zoomed to billions of users and multi-billion dollar values, and likewise, video sharing has exploded, like Vine and YouTube
  • chat apps — where discussions are the social objects — have likewise exploded.

In the enterprise side, we’ve seen the echo of Engestrom’s words: the best work management tools are those that focus on the tight connection between people in work networks, but take some aspect of the objects of interest floating around in the workplace — the things that define or refine our working together — and tightly connect them into a work graph. The best examples of all work management categories do this, like Asana’s focus on tasks, Podio’s focus on datasets (‘apps’ in their terminology), or Atlassian’s focus on the objects of interest to developers.

This is one of the factors that suggest why the best tools are tightly focused on a small number of work objects and the relations between those objects and the people that connect through them.

Justin Rosenstein of Asana recently talked about this in a Wired piece, The Way We Work Is Soul-Sucking, But Social Networks Are Not the Fix:

A work graph consists of the units of work (tasks, ideas, clients, goals, agenda items); information about that work (relevant conversations, files, status, metadata); how it all fits together; and then the people involved with the work (who’s responsible for what? which people need to be kept in the loop?).

The upshot of the latter data structure is having all the information we need when we need it. Where the enterprise social graph requires blasting a whole team with messages like “Hey, has anyone started working on this yet?”, we can just query the work graph and efficiently find out exactly who’s working on that task and how much progress they’ve made. Where the enterprise social graph model depends on serendipity, the work graph model routes information with purpose: towards driving projects to conclusions.

In my jargon, Justin is saying that push-oriented enterprise social networks (or ‘work media’ tools) are not the solution to work productivity, any more than email is.

My sense is that the reason we are seeing a stall in the uptake of the current generation of work media apps (enterprise social networks, social ‘collaboration’ tools, etc.) is that they don’t stick close enough to the work graph and pull communications , and focus too much on the network and push communications.

Don’t get me wrong. I am not saying all we need is a shared file system and a way to chat. On the contrary. But we have to get the dynamics right. When people are talking about work, or sharing work objects, the objects must almost be treated as people too, with deep metadata, persistent identities, and following/follower relationship with other objects and people in the graph.

In the new Pew Internet report on the Internet of Things in 2025 I was cited for inventing the term ‘computication’:

Desktop computers will be in museums, although a certain cadre will not give up their keyboards and will resist learning how to subvocalize or sign. People who talk to their goggles are considered infantile, since most people give up on that technique before starting school. Most people have wrist or finger devices that talk with their goggles, even when the goggles are off (in bed, exercising, in the shower, etc.), giving notifications, and allowing a subset of computication capability.

Some work objects — and other social objects — will become partly animate, capable of communicating with each other and us. Sensors, AIs, intelligent documents — all will demonstrate the characteristics of Bruce Sterling’s spimes. They will have ‘lifetimes’ and they will persist. They will have relationships with us and other spimes. They will computicate.

Our work graphs will be richer for that, but today’s tools are organized around inanimate and flimsy work objects. Beefing that up is one of the major trends for the next five years in the enterprise.