This is Going to Take Some Work...

Making Social CRM “Friend Discovery” Real

This is Going to Take Some Work...“Friend Discovery” is a new idea for Social CRM that connects the Customer Relationship Management (CRM) database in an organization with the social networks of its customers and other stakeholders.

This post goes into (even more) detail than my original one, “How Social CRM ‘Friend Discovery’ will Revolutionize the Way Organizations Collaborate.” You’ll want to read that one and possibly the “3 Principles of Friend Discovery Social CRM“ before reading this.

A Quick Recap:

Friend Discovery is the Intersection of CRM and the Social Graph

The first thing to note about Friend Discovery is that it doesn’t actually exist yet; at this point, it’s still just an idea.

Friend Discovery enables people to instantly discover friends who share an interest in some aspect of an organization. Friend Discovery goes much deeper than simply “liking” a webpage; it connects with the internal databases of the organization, and in so doing, enables me to collaborate with my friends in some very powerful ways.

A song by a new artist on Rhapsody becomes a lot more interesting when I see that a handful of friends are listening to it over and over again. An architectural firm is more credible when recommendations from people I trust instantly appear when I visit the firm’s website. Friend Discovery gives me rich ways to collaborate with friends around these recommendations and songs. It connects my social graph with the databases inside these organizations, not just the pages on their websites.

How Friend Discovery Works

One way to explain Friend Discovery is to look at what’s happening in the blue, green and yellow sections of the above Venn diagram – the intersection between the organization’s CRM and the customer’s social graph.

We can even reuse the more specific example of the iPhone upgrade scenario from my first article. In that example, I visited the iPhone page on and instantly saw all my friends who’ve upgraded from the same phone I currently have, which enables me to quickly and easily confer with those friends on whether they thought the upgrade was worthwhile.

Blue: The Organization & Its CRM

Apple has a CRM database to track customers and their purchases. Apple knows I have an older, 3G version of the iPhone and that I also have an iPad. They also know what products my friends have.

With a CRM database like, companies like Apple can easily and safely share slices of their CRM data using services like Customer Portal and Siteforce. Tools like this allow Apple to share information with customers like me without exposing it in ways that other people besides me can see. That’s a critical first step to Friend Discovery.


Now imagine if Apple were to use these tools to show me a list of all of their customers who had the phone that I’m thinking about buying; maybe even adding an asterisk next to people who upgraded from the exact same phone that I currently have. This is easy to do with today’s technology, but it raises some problems.

The first problem is that exposing this information violates the first principle of Friend Discovery, since those other customers probably don’t want that information out there for all the world to see. And even if they could limit who sees that information to just their friends, they’d probably be ok with sharing some but not all of their information with friends. In this case, I might let Apple share my music tastes from iTunes, and my ownership of various Apple electronics, but maybe not my interactions with Apple support. For Mutuality to work, the CRM needs to let me set permissions on what information I’m willing to share.

The second problem is that showing me all of Apple’s iPhone customers not only violates the third principle of Friend Discovery, it’s not even all that useful to me as a customer. I wouldn’t know what to do with all those names even if I could see them. What I need is a way to easily filter those customer records down to just the people I know.

Green: “Mutuality” Escrow Services

Mutuality Services

“Mutuality” is the way you do that; it’s the intersection of overlapping data sets, like the green “my friends with iPhone upgrades” slice in the above Venn diagram.

If you and I are both friends with Sara, we can both say she’s a mutual friend. Friend Discovery is similar in concept, except that it’s the friends I share with an organization. If you and I are both customers of Apple’s latest iPhone, we can both say we have a mutual interest in that Apple iPhone. We get at that overlap, or mutuality by comparing the people in Apple’s CRM with the friends in my social graph.

One of the key technical challenges of Friend Discovery is comparing the organization’s and its stakeholders’ relationship data in ways that don’t compromise it. Apple doesn’t want me snooping around in or stealing information from its CRM database and I don’t want it doing the same in my social graph.

“Mutuality escrow services” are how that happens. Mutuality escrow is a third-party service that takes a temporary “snapshot” of the customer’s social graph and the company’s CRM in order to calculate the overlapping records. Once that “mutual dataset” is determined, the escrow provider immediately deletes the non-overlapping records. The overlapping “mutuality dataset” is then passed to a secure – and isolated – set of database tables connected to the organization’s CRM. The tables are sequestered from the rest of the CRM data to prevent the company from retaining their customers’ social graph information (note: this would probably need to be reinforced by clauses in the organization’s privacy policy as well). That mutuality data is then used to drive Friend Discovery applications on the organization’s website (such as those highlighted in the first Friend Discovery article).

Mutuality escrow represents a new business opportunity for companies experienced in processing large data sets. Mutuality data sets will be quite large and need to be calculated in real-time. That combination of scale and speed makes it critical that organizations have reliable mappings between customer records in their CRM and their stakeholders’ social graphs. The easiest way to do that mapping is to have customers provide their online social identity as part of their initial account sign-up process, which they will be more motivated to do when they understand the benefits of the organization’s Friend Discovery applications.

Yellow: The Customer’s Social Graph

The Social Graph

The social graph presents some real challenges to Friend Discovery. Today, I don’t have one common identity that I use online. Instead, I have identities on Facebook, Twitter, Quora, Yelp and lots of other services, each of which connects me with different slices of my total social graph. Until we find a way to synchronize these disparate social graphs and identities, Friend Discovery is going to be difficult.

One solution to the fragmented social graph is, of course, for one company to control all of our online identities and social graphs. Some would say this is a reality that’s rapidly emerging with Facebook’s runaway success. Personally, I really like Facebook. I admire the polish and ease-of-use of the service and use it a lot. But online identity and the social graph are emerging as an essential element of modern society; allowing one company to control this societal function creates a host of problems. Friend Discovery shines a light on a few of them.

I trust some of my Facebook friends more than others, and there are quite a few of my Twitter followers who I don’t really know at all. There are lots of reasons why I may not want everyone in my social graph knowing that I just checked in to the The Mirage in Las Vegas or that I just bought a fancy new LED TV that’s now sitting all by itself in my home right now. This is the second principle of Friend Discovery which is that I need to be able to tier my social graph based on trust.

Here’s one of the problems with one all-powerful social graph-controlling company: let’s say the company didn’t believe this trust-tiering thing was a real priority, or that they simply implemented it in a way I don’t like. If they’re the only game in town, I’m out of luck; another company isn’t going to come in and beat Facebook, for example, just because of this trust-tiering feature. I have a lot of time and energy sunk into my social graph on Facebook and the benefits of this feature don’t overcome the costs of my having to completely reconstruct my social graph someplace else. That lack of openness in social graph solutions reduces our control over our social graph, which leads to less competition, less innovation and less diversity of solutions.

And that turns out to be a real problem when it comes to Friend Discovery.

Technological innovations bear the birth marks of their origins. To succeed, consumer technology must optimize for simplicity and easy-of-use. Business technology, on the other hand, needs to deal with complexity and optimize for flexibility, even if that means some sacrifice in simplicity. Friend Discovery is the contact zone between consumer technology (the social graph) and business technology (the CRM). We each have hundreds, if not thousands of relationships with businesses and other types of organizations in our lives. Our interactions with those organizations are inherently quite diverse and, in many cases, quite complex. Any solution for Friend Discovery that forces consumer-level simplicity on this real world complexity and diversity is bound to fall short of its true potential. Sure, good design enables some degree of both simplicity and flexibility – but in Friend Discovery, flexibility reigns.

Facebook is opening its social graph to enable users to tag a limited number of real world objects so that they show up in Facebook users’ social graph and activity streams. But is Facebook really the entity best suited to mapping the vast variety of real-world objects and infinite ways we use those objects to connect with the incredibly diverse organizations and people in our lives? I have serious doubts.

Finally, the third principle of Friend Discovery focuses on ensuring that relationship data stays where it belongs. Customer relationships belong in the CRM and friend relationships in the social graph. If Facebook does emerge as the centralized point of control over people’s social graphs, that poses a big threat to all organizations interested in Friend Discovery services. Apple, for example, is not going to want a powerful company like Facebook using its control over the social graph to reverse-engineer Apple’s customer base – something that is entirely possible with the existing technology. “Ping” is one logical response to this concern.


So, what’s the right solution for the social graph? Two words come immediately to mind: open and federated.

There are a number of smart people thinking and setting good principles around a more open social graph. Summarizing their ideas is beyond the scope of this post, but it’s something we all need to care about – and not just if we want to see Friend Discovery realized.

As far as federation is concerned, I really like the model that Automattic has built with WordPress. They run their own hosted version, but also allow users to run WordPress software on third-party hosting sites. You could imagine Google opening up Profiles in such a way that users got to choose whether to keep their identity with Google or host it with third-party i-brokers. Augment these profiles with the social graph, possibly with a kick start from GMail, and you’ve got a good foundation.

The key of course, is convincing users that it’s worth investing time in building and curating a new, more open version of their social graph. I believe that Friend Discovery could provide just such an incentive.

Imagine if Netflix, Amazon, Yelp, Groupon and other big online sites with a stake in the future of the social graph were suddenly to implement some really cool Friend Discovery applications based on a new, open social graph. Now, as a customer, I can really do some interesting coordinating with friends on movie picks, favorite authors, local deals and so on. Not surface-level liking of pages, but deep interactions with the valuable information contained in these companies’ databases.

CRM solution providers like Salesforce would also have huge reason to get involved as a way of extending these features to their CRM customers – organizations that could also benefit enormously from Friend Discovery services.

Friend Discovery is the next killer application of the social graph, and I believe it’s big enough to blow it wide open.


[blackbirdpie id=”53252728893358080″]

Follow Gideon on Twitter: @gideonro



  1. Good article. Many aspects of what you touched on reminded me of Watsa on Facebook. The connector of course would be a CRM like SFDC. I could see convergence in those three products and concepts to deliver some of the theories you have talked about. Good stuff!

    • Geoff – thanks for the note. Sorry for the delayed response. Just catching up after vacation. I am not familiar with Watsa and couldn’t find a reference to it. Can you provide a pointer?