Explaining what the “Social Graph” is to your Executives

Categories: Social Graph, Web Strategy, Web TheoryPosted on November 10th, 2007

Many are talking about the Social Graph at the technical level, leaving many business folks with many questions. The following, using clear business language (void of tech speak), will explain what is the Social Graph, why it matters, and what you should do.

In respecting your time, I’ve boiled down the entire post to seventy words:

Executive Summary
The Social Graph is the representation of our relationships. Today, these graphs define our personal, family, or business communities on social websites. Unfortunately, we’re duplicating our same Social Graph on multiple websites, resulting in inaccurate data and time spent managing it. Despite many challenges, our Social Graphs should be self-managed from a single trusted source, replicated to websites of our choosing, thus resulting in accurate, efficient, relationship management.

Want to be able to have an intelligent conversation with others on the topic? The detailed, more in-depth analysis follows:

Situation
Relationships are nothing new, from paper based address books, to the arrangement of huts in early villages, we’re social creatures. The big change? The internet has broken down physical barriers, allowing us to connect and create online communities.

Online social networks such as LinkedIn, Facebook, Ancestry.com, and thousands of others are increasing in global adoption. I predict that social networking features will be part of nearly every website.

While we have many networks, and thus many social graphs they often fit into three major categories: business, personal, and family. Alex Iksold breaks down the components to three specific elements: People Identity, Type of Relationships, and Relationships Identity. These relationships can be mapped, displayed, or documented in what we call the Social Graph.

[Definition: The Social Graph is the representation of our relationships. In present day context, these graphs define our personal, family, or business communities on social networking websites]

Pains
Today, individuals are part of many online social networks. Sadly, the same Social Graph is inefficiently being duplicated; we’re forced to add the same contacts to each social network. Beyond frustration, this results in inaccurate networks, and inability to control our own data. As we join and register more websites, this pain will grow exponentially.

[Unfortunately, we’re duplicating our same Social Graph information on websites with social features, it's annoying to manage and inaccurate across sites. These disparate and disjointed social graphs will not scale, ultimatly resulting in a fragmented user experience as we use multiple websites]

Opportunity: Aggregating Social Graphs
The idea is simple: Allow relationships to be quickly shared once and then replicated across multiple websites. A movement has been started to allow these relationships to be transplanted from one social network to another. The goal? reduce inefficient adding of relationships, improving the accuracy of the network, and providing users with control and management of their relationship data.

Benefits

Users: For users this means efficiency and control over one’s personal data, their relationships, and how they are deployed on different social networks, it makes navigating the web better.

Social Networks: For companies that are social networks, they can benefit by increasing the amount of users as the social graph will populate all of a users network they permit.

All other websites: For companies that are not currently social networks, (like a corporate website) expect these social features to be part of your site. People will co-surf and share information about your content whether you like it or not.

Scenario: What does success look like?
In the most ideal scenario, aggregation of the social graph would allow Ted to quickly move from one social network to another, his trusted contacts would be pre-populated (as specified by Ted) and sharing of information would be instant. Furthermore, Ted would be able to quickly aggregate all information about his friend Chris on one page, and see what Chris has chosen to share across all social networks. Lastly, Ted would have full control over his graphs and can remove, adjust, or modify at will. Ted is in control, has accurately updated information from his social networks, and saves time at work, in his personal life, and with his family. He has more time to do what he really wants, which is to read this Web Strategy blog.

How can this be accomplished?
By centralizing a users Social Graph on a trusted, third party area that can be a central place where relationships are updated, and then replicated to every social networking website using a common process and technology.

Many Challenges
Like many new initiatives, we’ve not laid out the infrastructure for open relationship movement, there’s a few obstacles to overcome.

Social Network vendors scared to open up and let customers and their relationships easily move to other networks

Agreement needed between all vendors and participants

Ownership over project and data

Lack of general market awareness

User adoption (sadly, I think most users are sheep)

Likely, a need for a single login

Creation and costs of third-party silo

Privacy concerns: many European countries may not embrace

Multiple security issues

Legal and government may get involved

As an industry, these challenges are not insurmountable, what’s required is a clear roadmap, willingness for vendors to ‘let go’ and industry-wide support.

Future Outlook
Recently, an alliance called OpenSocial was formed between Google, MySpace, (list) to create an agreement so applications could easily be ported from one network to another. Although still at it’s starting point this alliance should be the precursor to establishing the management and flexibility of the Social Graph. To learn more about OpenSocial read: Explaining OpenSocial to your executives.

[Ideally, our Social Graphs should be self-managed from a single trusted source, and then replicated to websites of our choosing, thus resulting in accurate, efficient relationship management]

What you should do:

1) Wait and Watch: We’re still at the conceptual level, although several vendors are working hard to release products or services to fix this, there’s not much to do.

2) Plan: If you work for a company that has relationships on it’s website (having a ‘login’ is a key indicator) then you should make your web teams aware of this market change, and have an internal discussion.

3) Obtain more information: I’ve created a new tag called Social Graph on this blog, and will keep you updated of changes as I see it. Being an industry analyst, vendors will brief me, I’ll analyze with my peers, and report back.

Additional Resources:

Brad Fitzpatrick, who’s leading the movement provides solutions and discusses where he’s at, developers should work with him.
Plaxo, a vendor in this space is moving forward: Building an Open Social Graph

Six Apart: Is supporting and will release versions of the social graph

Alex Iksold: Social Graph Concepts and Issues

Jim Calhoun: Has a great video where he explains how the Social Graph works

Josh of RWW suggests that we should retire the term Social Graph and use the term Social Network

Dave Winer: Suggest we use the term social network, I get what he’s saying, yet the need for portability and aggregation between social networks that is the crux of the issue.

I suggest that one way is to use a browser as a viewer for the social graph

Why this analysis?
When I’m asked the same question several times (in this case, colleague Jennifer D and Bill Claxton), it’s a signal to write an explanatory blog post.

Feedback from peers
I’m trying to be as accurate as possible and incorporated the feedback of the two-line definition from my peers at the Blog World Expo which included Marshall Kirkpatrick, Jeremy Pepper, Jake McKee, Stephanie Agresta, Chris Heuer, and anyone else who would listen to me. Please note they did not approve or review the rest of the text.

Your comments wanted, even if you don’t agree
For more posts like this, I’ve tagged it Web Strategy, where I publish many how-to, and what-you-should-know posts

If you have suggestions to improve this document, leave a comment, and I’ll update as appropriate, I hope you share this with others.

Share This Post:

This entry was posted on Saturday, November 10th, 2007 at 8:33 am and is filed under Social Graph, Web Strategy, Web Theory. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

Comments

ameer

Fine, thank you for your clarification and good ideas.

Cheers, :)
Ameer.

Scott Germaise

"Unfortunately, we’re duplicating our same Social Graph on multiple websites, resulting in inaccurate data and time spent managing it."

We're actually doing a lot more - and less - than that. In a variety of cases, we're extending our social graph in ways we wouldn't have before. Sometimes cluttering it up with unnecessary, tenuous or not quite real relationships.

"User adoption (sadly, I think most users are sheep)"

Maybe. Or maybe most users are busy. This seems to come up as a theme almost again and again in our industry. People who build some of these tools - including me - can't forget that the mechanisms don't matter so much as the value. Most users don't work in the Internet industry. Regardless of how much time Gen Y or Millenials, or whatever consumer segment may be adopting these products, most folks are doing other things. Building buildings, saving lives, serving up fries, whatever, etc. They're not obsessing over the best ways to manage their personal info online. If anything, I think it's conceivable the "learned helplessness" regarding control of one's information has most folks just either not caring or tuning out to it. There are, after all, only so many places you really need your social graph to be anyway.

Scott

Jim Goldstein

Great post. I'll definitely be referencing this for others I communicate with on the subject.

I do find the promise of a referenceable network resource interesting but it would seem to be a double-edged sword. For both business owners and users alike the "flavor of the month" phenomenon for web sites/services is a real distraction/detriment. From the business side it poses the risk that you might lose customers even if you take positive steps in developing your business all because your site isn't the "in" site. Conversely if you're a user you might be chasing friends from one site/service to another all because of what is considered in fashion. I see this all the time now Twitter to Pownce to Jaiku to Twitter to Facebook and so on. At some point fatigue sets in turning off users from adopting new and potentially better services.

Being able to replicate the social graph between sites/services on the surface seems to be a great idea, but it leaves me thinking perhaps it is a patchwork fix to a social dynamic issue that isn't fully studied or understood. Time will tell I suppose.

andrey

great summary, man. as usual you are quicker than anyone else! for what is concerning third party, neutral place to store the graphs, i think a service like spock would be best. But i am not a dreamer- Gooogle will come soon and eat up it all. it's for sure their strategy- to generate some great interest from this new idea, and then to make it part of Google... but let's hope it will be a neutral place ;-)

Jeanette

Jeremiah,

You're funny, too! "SHe has more time to do what she really wants, which is to read this Web Strategy blog."

I have noticed that many of my Linkedin friends won't join Facebook. In fact, my friends on Facebook are totally different than my Linkedin connections.

What would be the best benefit to me--the ability to send one message to everyone.

Thank you for continuing this discussion. I would have missed it if not for your twitter.

Joy!
Jeanette Fisher

John Stack

I can't believe people wouldn't attach this type of ontology to sites. Of course you have readers and they respond in different ways!
I first dug into this with NVU, and I thought - fascinating - now, only to be able to match content to readership/target market.

Can anyone tell me how my site woud like up socially? http://epaper.hubspot.com

jeremiah_owyang

Ralf

This is the spec that Brad Fitzpatrick wrote out in his design spec --that it be on a third party trusted source.

Ralf

My basic concept of "social graph":

If I have a relation to a person in one social network all I have really is a pointer to another persons node in that network. Because I have to more than one person a relation I have several pointers to other nodes in that network. Because different social networks offer different kind of value I have to maintain my "I" nodes in all of them.
If I would draw those pointers with the "I" node in the middle and the other nodes (my contacts) around "I" I would get a "star". This star is what my paper address book can also represent well. A social graph is "more" than that. A social graph is the combination of these "stars" for many "I"'s.

A provider of a social network is offering to users the functionality to setup and maintain their "I" node and to maintain their set of pointers. Interesting functionality is offered to users (and the industry) on the basis of the social graph that is stored and maintained in the underlying database of the social network. In this sense with today's immature status of the social network industry something called a "social network" is in the hands of the people that own the existing social networks.

Introducing the concept of virtual nodes and attributes of pointers:

A pointer can also point to a node that does not represent another person but instead it would be a virtual node that other people can point also to. This virtual node could be used in a specific network (implemented in the application layer of the social network; the social graph in this social network is its database/structure) to implement for example the concept of a "Group". Another usage scenarios example could be the concept "meeting" taking time as an attribute for relations into consideration.
Other attributes can be attached to relations between the nodes. In the context of representing relations between humans "family", "friends" and "business" are common high level attributes to segment relations. On a more detailed level "is mother of", "is boss", "worked at the same company" etc. are common attributes that specialised networks are offering.

Bottom line:

You are saying in this interesting article: "Ideally, our Social Graphs should be self-managed from a single trusted source, and then replicated to websites of our choosing, thus resulting in accurate, efficient relationship management"

Based on the different usage scenarios of virtual nodes and various types of attributes I am not sure that a single trusted source is the ideal or even a feasible solution.
Looking at the basic mechanism of the web -the URL- and its incredible value maybe someone can come up with a secure solution where I just "point" to my already existing nodes in other social networks and this new solution would not import/copy but instead directly leverage the star representation already existing in the other networks (maybe this can be achieved with the API Google suggested and if not maybe with an enhanced version?). Based on that this new social network could offer its great new functionality and value for me as the user.

Thank you for your article.

jeremiah_owyang

Mitchel

Yup, I JUST heard about this category recently, but doesn't this apply under business/work?

Mitchel Ahern

Hey Jeremiah,

I'd like to propose that in addition to business, personal, and family as major categories you might consider adding "artist" as well. Now I know this could either be considered either a flavor of personal or at best a "minor" category, and yet the nature of the net and today's technology means that a surprisingly large number of people are out there with their own creative endeavors, which function in some ways like a business, but wholly removed from their normal business channels.

Of course the best evidence for the strength of this is how MySpace handles musician accounts as wholly different thing than their regular accounts.

Mitchel Ahern

Ralf

My conceptual view of "social graph":

"I" as a single person participating in a social network do not have or own one or more social graphs.

Basic concept:

If I have a relation to a person in one social network all I have really is a pointer to another persons node in that network. Because I have to more than one person a relation I have several pointers to other nodes in that network. Because different social networks offer different kind of value I have to maintain my "I" nodes in all of them.
If I would draw those pointers with the "I" node in the middle and the other nodes (my contacts) around "I" I would get a "star". This star is what my paper address book can also represent well. A social graph is "more" than that. A social graph is the combination of these "stars" for many "I"'s.

A provider of a social network is offering to users the functionality to setup and maintain their "I" node and to maintain their set of pointers. Interesting functionality is offered to users (and the industry) on the basis of the social graph that is stored and maintained in the underlying database of the social network. In this sense with today's immature status of the social network industry something called a "social network" is in the hands of the people that own the existing social networks.

Introducing the concept of virtual nodes and attributes of pointers:

A pointer can also point to a node that does not represent another person but instead it would be a virtual node that other people can point also to. This virtual node could be used in a specific network (implemented in the application layer of the social network; the social graph in this social network is its database/structure) to implement for example the concept of a "Group". Another usage scenarios example could be the concept "meeting" taking time as an attribute for relations into consideration.
Other attributes can be attached to relations between the nodes. In the context of representing relations between humans "family", "friends" and "business" are common high level attributes to segment relations. On a more detailed level "is mother of", "is boss", "worked at the same company" etc. are common attributes that specialised networks are offering.

Bottom line:

You are saying in this interesting article: "Ideally, our Social Graphs should be self-managed from a single trusted source, and then replicated to websites of our choosing, thus resulting in accurate, efficient relationship management"

Based on the different usage scenarios of virtual nodes and various types of attributes I am not sure that a single trusted source is the ideal or even a feasible solution.
Looking at the basic mechanism of the web -the URL- and its incredible value maybe someone can come up with a secure solution where I just "point" to my already existing nodes in other social networks and this new solution would not import/copy but instead directly leverage the star representation already existing in the other networks (maybe this can be achieved with the API Google suggested and if not maybe with an enhanced version?). Based on that this new social network could offer its great new functionality and value for me as the user.

Thank you for your article.

Bill Claxton

Thanks for this clarification Jeremiah. Spot on as usual, and I appreciate you advancing the discussion by zooming in on the inefficiency of having this graph data replicated in several different sites.

This topic is beginning to resonate with the idea of 'presence', as in, how is my online presence represented in different contexts.

I'm reading a great book titled 'Crowd of One' that is providing me lots of insights on social currency. Not finished yet, but I recommend it to all interested in the area of online identity, especially the issue of how we build trust and use social currency online.

jeremiah_owyang

Len,

If your company already has several 'social features' on different websites (logins, with profiles that let people connect to each other are a clue) then you move into the social networking category.

A big company, like EMC, that focuses on organizing information, storing and managing it, should consider being part of this leadership of the aggregated social graph, likely enterprises are going to need a solution to integrate external and internal social graphs.

Back to your original question, as those companies build micro-silos, at the very least, see if the identity information and profile information can use a single authentication system.

Looking forward, I'm expecting a third-party 'social graph solution provider' to help fix these problems.

Len Devanna

Good information, as usual.

A question to one of your comments - "There’s nothing for companies that are not social networks to do at this point."

What about the company who has several different initiatives all trending to becoming independent micro social networks (Each being managed by different business units such that there's little cohesion). While I agree we're in 'wait and see' more, I wonder what can be done pro actively to minimize the eventual pain of adoption?

David Cushman

Hi Jeremiah:
Thanks for enabling the conversation on this.

Here's my contribution: http://fasterfuture.blogspot.com/2007/11/social...

jeremiah_owyang

Carlos thanks, but no reason to 'Yell' I hear ya!

carlos zucchini

WE COME A LONG WAYS IN DEVELOPING APPLICATION AROUND SINGLE COMPANY, WHICH TAKE THE NASDA UP TO NEW HIGHTS THEN TO DISCOVER THE SINGLE APPLICATION NOW ARE APART OF BRANDED [SERACH] SITES OR COMMUNITIES. SINCE WE FOGET THIS WAS BORN OF THE COMMUNITY SO NOT BE SUPRISE THE RETURN TO THE COMMUNITY AFTER YOUR TRIP AROUND THE WORLD. HAVING ONE PAGE TO VIEW MULTIPLE COMMUNITY IS NOT WITHOUT MERIT BUT TO HAVE ONE CENTRAL [AREA] TO HOUSE THIS INFO. WHEN COST IS REDUCING AND SECURITY IS FIRST IN OPERATORS AND MEMBERS MIND MIGHT PROVE TO BE DIFFICULT TO SELL. OPENSOCIAL IS JUST THE START WHERE MULTIPLE COMMUNITIES WILL EXIST...using newly improve application or one off the shelves. all we are doing here is CREATING A LARGER MORE RELIABLE PLATFORM FOR ADVERTISEMENT DISTRIBUTION. USING CREATIVE IDEAS PLUS LEVERAGING APPLICATION TO GIVE THE USER A BETTER EXPERIENCE IS FIRT..INDIVIDUAL WILL BUY UNDER ANY PLEASANT CIRCUMSTANCES.

jeremiah_owyang

Michael

I vetted this summary a few times, putting myself in the situation of talking to a CEO, telling him/her only what they need to know.

I read your previous Tweet:

1) "Read your Social Graph exec. summary, and don't get it. Where's my social graph, and why does someone else need to manage it?"
I answered where the social graph on line 3 of the exec summary: "Unfortunately, we’re duplicating our same Social Graph on multiple websites"

2) "and why does someone else need to manage it?" I think that was also answered in the exec summary: "...our Social Graphs should be self-managed from a single trusted source, replicated to websites of our choosing, thus resulting in accurate, efficient, relationship management."

To those hairs, How would you restructure the executive summary? What elements are missing? What else should an executive know?

Will make changes based on feedback, thanks Michael, I'm listening and would love to make this better.

Michael Krigsman

Based on your Twitter announcement, I was looking forward to a clear, business explanation of Social Graph in the executive summary. The executive summary doesn't deliver on that promise.

In effect, you actually established two goals in the executive summary: a. clear presentation, and b. short word count. Unfortunately, the word count "writing challenge" overpowered the more important communication goal.

Perhaps it may appear I'm splitting hairs, but isn't that what we bloggers do? Of course, feel free to do the same over at my ZDNet blog. Critical comments always welcome!

Cheers,

Michael Krigsman
http://blogs.zdnet.com/projectfailures

jeremiah_owyang

DJ

Check out this Facebook group I created called "The Web Strategy Group" it's really an extension of this blog, except the discussions can be created and started by you and others (I do enough talking and writing)

This conversation about the 'socialization of corporate websites' seems so relevant:

http://www.facebook.com/topic.php?uid=259158268...

jeremiah_owyang

DJ

In the "benefits" section I segmented the three different types of websites.

There's nothing for companies that are not social networks to do at this point.

Expect the aggregation of Social Graphs to impact companies that are participating the OpenSocial project (Google, Myspace, Linkedin, Oracle, SalesForce, Plaxo, SixPart and others). Please note they all have Social network components to their sites.

After it's deployed to these first test companies, we'll start to see some widgets appear that can be embedded on other websites (like static corporate ones ...of no mention)

This is going to take some time, months, if not by end of 2008.

The next steps I prescribed still apply: Wait and Watch, Plan (if your site has a login), and keep on learning by coming back to this website.

I'll guide you using my blog. I'll tag relevant posts with "socialgraph".

And yes, I attended three rigorous weeks of Forrester training, and am continually being challenged how I think, how I communicate, and the statements I make to be backed up with reason, fact, and research. Thanks for noticing!

DJHowatt

Jeremiah,
The Forrester discipline is coming through in your insights. Great work.

So how does a company -- no names mentioned -- take the first step? We can't go to full-on support of social graphs until our own infrastructure catches up. Is a baby step the adding of "Technorati Links • Save to del.icio.us • Digg This! • Email this • Stumble It!" to our pages and press releases? Something else?

jeremiah_owyang

Sean, I'd love to read your additional thoughts, do come back and leave a comment.

NMW, Yup, the way I see it, there are multiple social graphs within each of the three major buckets.

nmw

This reminds me of the expression:

"don't confuse the map with the territory"

Just as there are different maps (e.g. topographic, political, satellite image, radar, etc.), it makes perfect sense that there will be different SOCIAL GRAPHS, not just a "one size fits all" SOCIAL GRAPH.

For example, take the aspect known as "religion". One GRAPH might simply be a binary bit "religious" (yes/no). Another GRAPH might list 4 or 5 "major religions" and ask to "select one". Another GRAPH might list hundreds of religious views and ask to mark all that apply.

If there were a social network that was focused on religion, I would expect this GRAPH to be quite detailed. In a social network focused on sports, it might not be significant.

Each network will gather the data its community members feel are important to their collaborative project.

Sean O'Driscoll

Very useful discussion - I like the way you've top lined this. I have not blogged this yet, but next week I'm delivering a presentation on influencer program development and management at the WOMMA conference (see post here: http://communitygrouptherapy.com/2007/10/20/wil...). At the close of the presentation I try to anchor the model and next steps back to the social graph. In this, I've been working through three layers of descriptions - as ultimately you have to make this not just acedemic but actionable for (in this case) marketers. Part one I simply describe as the social "map" - which is the relationships between everyone in a network - the topology. I then describe the Social graph as the meta-data or properties that describe the relationships between people (how far, how fluid, etc). The point of this is that if I whisper to 2 different people in a network the information will flow through at different rates. Understanding what governs the rate (and accuracy) of how information flows implies a deeper understanding of the relationships than that person a knows person b. The 3rd level of this that I've been thinking through is the notion of Personas in a social graph - Conversation starters, stoppers, branchers, trolls, accelorators, inactives, etc. I introduce all of this as ultimately the topic is about influencers in the network. Depending on your business objectives, you should be able to define the persona and characteristics of the influencers in the network you need to reach - to aid in the process of influencer detection and engagement. For example, as a marketer, i may be uniquely interested in alpha talkers/conversation starters. As a support exec, I may not care about conversation starters...I want to know the conversation stoppers - those that are answering all the questions in my support forums.

i'll have to blog this more next week after the talk. I'd love your feedback on the approach.

sean

Brandon

Great post! Definitely a well articulated discussion of a topic with a lot of buzz and many descriptors.

jeremiah_owyang

Mike

ABSOLUTELY, Share it with others, that's why I wrote it!

MIke Coulter

Wow, Jeremiah,

What a great overview, (if indeed anyone can call the difficult task/trick of encompassing the Social Graph into 70 words, an 'overview'.)

I think this is a really important post at this stage of the web and the evolution/next stage of social networks.

Hope it's okay to forward your analysis to clients and colleagues; who are all still trying to figure out what you have articulated so well here.

Bravo.

jeremiah_owyang

@Shawn Zehnder Lea

1) Yes, that's assumed you can take other social information with you.

2) Great analogy, but I don't think it applies to all of the globe.

3) It's inaccurate because you may have a business network on XING and LINKEDIN, yet the contacts are all different. If Tom is a friend of yours in one network (of the same type, say business) then you shouldn't have to specify he's a friend in another network. It should be automatic.

jeremiah_owyang

Shawn has some great public comments from Twitter, I've added them here:

Shawn Zehnder Lea:

@jowyang ...we should be able to take our social information too

@jowyang You could use the cell phone analogy - we now believe in the right to take our personal phone numbers with us...

@jowyang How is duplicating our own social graph creating inaccurate data if we are doing the duplicating ourselves? User error?

Peter Bihr

Jeremiah, great stuff! You might also want to check out that video: We asked Six Apart's David Recordon how he would explain the Social Graph to his parents, here's how (starting at 1:50).

Sorry the rest of the video is mostly in German and we haven't had the time to cut out David's statement, but if you jump about 1:50 into the video, you'll get a great non-technical explanation of what the Social Graph is.