Gartner Blog Network

Facebook Enlists Friend-Sourcing in Preemptive Strike Against Google

by Ray Valdes  |  October 8, 2010  |  2 Comments

The engineering team at Facebook has been in “lockdown” mode for the past 60 days, and yesterday released the results of their labors. The launch event at Facebook offices in Palo Alto unveiled a whole bunch of enhancements: an improved version of Groups, plus a data-export facility for your social data, as well as a better user administration dashboard. These improvements follow in the wake of recent work in high-resolution photos and social gaming support.

I managed to squeeze in some time to attend the event yesterday, given short notice amidst many other existing commitments. Today, I’m gathering my thoughts, not just about yesterday’s announcement, but about the big picture — one which has to do with a “worlds in collision” dynamic between Facebook and Google.

First, a bit of history. A key reason behind Facebook’s success has been their unrelenting drive, sparked by CEO Mark Zuckerberg, to evolve the site, in pursuit of a moving target: a highly refined but still evolving vision of the Social Web. The pursuit of this vision has not been smooth: mistakes have been made with regard to privacy, users have been regularly pushed to the edge of their comfort zone regarding modes of online interaction and information sharing, and the patience of developers has been regularly tested with regard to changes in rules of engagement (i.e., how apps can message users) and also tested by a shifting API that is only partially documented. Nevertheless, last month, Facebook’s growth trajectory passed the 500 million user milestone (active users, ones that log in on average at least every other day) — a trajectory that seemingly will continue indefinitely.

However, the spot on the top of the heap that is the Social Web is not a stable place, as staffers at former top-dog MySpace will attest. The latest threat to Facebook’s dominance comes from Google, which in past years has repeatedly tried and failed to get onboard the Social Web train (through past acquisitions such as Jaiku and Dodgeball, and internal projects such as Buzz and Wave). People at Google observe the macro-trend as the Web evolves from content-centric to people-centric, and have recently marshalled forces for a renewed assault on this new market territory. Google has acquired some companies, recruited outside talent, repositioned existing management, and, rumor has it, is set to unveil a major initiative in the Social Web.

Google fired a shot across the bow for Facebook 3 months ago, with the publication (on of a presentation by Paul Adams, a senior user-experience designer. This presentation got about 300,000 views, an achievement for a 224-slide deck on a specialized topic. The situation that Adams identified seems at first rather obvious: people don’t have a single group of friends — in the sense of a homogeneous mass. Instead, people have multiple independent groups of contacts at varying levels of affinity and intimacy: family, best friends, college chums, work colleagues, ex-spouses, sports clubs, etc. The one-line problem statement is self-evident, and this is something that has been known for a long time. However, the solution is quite elusive, and depends on careful analysis of the situation. Yesterday, I chatted briefly with Mark Zuckerberg about Adams’ work, and Zuckerberg did say that Adams presented a convincingly detailed and highly articulate statement of the problem. Zuckerberg also noted that Adams’ presentation did not offer any solution.

In case you are curious, here is a snippet of Adams’ analysis, showing how existing social networks like Facebook fall far short in meeting this basic requirement for an online “social utility”.

Facebook, before Groups

He calls out Facebook by name in this slide, in effect waving a red flag in front of cross-town rivals. The response, apparently, was to “lockdown” — to enter a mode where engineers are, um, strongly encouraged to work steadily and not leave the premises until the lockdown is lifted (in this case, 60 days).

At the launch event yesterday, there was no specific mention of Google or Adams’ work, but instead Zuckerberg talked about past efforts by Facebook at solving the problem of adding structure to the collection of friends. An early effort was Friends Lists, an oft-requested feature that Facebook introduced a while back but has seen little adoption (only 5% of users have created a list and usually stop at one or two lists). Zuckerberg also talked about the second step up in addressing this problem, the algorithmic approach, which Facebook software uses to prioritize and filter items on the news feed, and to show top contacts on the Chat list. Zuckerberg admitted that this was an imperfect process, and that when the software is wrong, it can sometimes be spectacularly wrong, such as showing your ex-spouse as your top contact. So the problem is long known, and solutions have been unsatisfying.

The solution Facebook unveiled, as a step beyond the manual solution and the algorithmic solution, is what they call the “social solution” — relying on a group of people to do a better job than one algorithm or a single individual working in isolation.

Unlike a Friends List, which is metadata (data about social data, i.e., one’s friends), a Facebook Group is crowdsourced by your friends. I call this approach “friend-sourcing”. (After I thought up this term, I was quite proud of myself, but then did a quick Google search and found that this is not new under the sun. Oh well, it is nevertheless apt.)

Groups avoids a limitation of Friends Lists, which are created by an individual and not visible to others. If Bob friends Alice, and Alice decides to put Bob in her private list of Annoying People (accompanied by reduced permission levels) that is obviously not something she wants Bob to see. And vice-versa, if Bob has a list called “Gals I Want to Date”, that is also something that the list’s creator does not want to share. The Friend List approach places the entire burden on the list creator, who works in isolation to create and manage social metadata. Facebook says only 5% of individuals bother, and users normally create only one or two such lists before feeling like it is too much work.

A Facebook Group (the new version, not the underpowered one that has been around for years), on other hand, is created by one person who identifies a collection of friends, and that work is visible to and shared by the others. Other people in that group benefit from that person’s effort, and can build on top of it. It is viral and, in Facebook’s approach, woven into the fabric of the Facebook social experience, meaning notifications appear in the News Feed, and one-to-one chat becomes Group chat.

The development project for Facebook Groups had at one time the codename “Tribes” — despite the California overtones, this term is quite accurate in describing the concept of what Adams calls groups that are bound with “strong ties”. The following slide by Adams has strong tribal overtones.

Paul Adams tribal-like illustration

I think Facebook Groups is a well-conceived and innovative solution to a long-standing and very real problem. In the past, the workaround for this problem (one that I use, along with some of my friends), has been to use multiple Facebook accounts (one for family, one for professional contacts, etc). This is a cumbersome and inefficient attempt to establish firewalls around pools of social data. Groups seems quite natural, by contrast.

Facebook Groups has gotten off to a robust start. It went live yesterday and I’ve noted a burst of activity in various circles. By itself, it is not a game changer, but it definitely solidifies Facebook’s defense against a possible assault from Google or from other would-be players in the social arena (Yahoo, Microsoft, Twitter). Likely there will be some privacy-related issues that surface around Groups, but these probably won’t be widespread. One example was raised by Jason Calacanis, who has many Facebook “friends”, one of whom joined Calacanis into a fake group with a name that many find offensive. This prank is mean but likely not typical. The solution is to defriend those people who play mean pranks on you.

The friend-sourcing approach to solving a design problem makes a lot of sense. It is a textbook example in how problems in social experience design can often have a socially-based solution. Facebook’s crowd-sourcing of language translation (a couple of years ago) is a precursor example.

How about you, how would you apply friend-sourcing? On a separate note, is there any issue that would compel your colleagues at work to enter a 60-day lockdown?

Additional Resources

View Free, Relevant Gartner Research

Gartner's research helps you cut through the complexity and deliver the knowledge you need to make the right decisions quickly, and with confidence.

Read Free Gartner Research


Ray Valdes
Research VP
9 years at Gartner
30 years IT industry

Ray Valdes is research director in Gartner Research, where he is part of the Internet Platforms and Web Services team. Read Full Bio

Thoughts on Facebook Enlists Friend-Sourcing in Preemptive Strike Against Google

  1. The idea of different groups of friends is something that has kept me from using Facebook as much as I wanted to in the past. For example I want to sometimes post a status or make a comment about something political, but I don’t want everybody on my Friends list to see it. There is no way that I know about to do that with Facebook right now. At least not in an easy way so they need to work on that. Maybe having Facebook choose which groups you can broadcast messages to on the fly would help?

  2. […] Ray Valdes is research director in Gartner Research, where he is part of the Internet Platforms and Web Services team. Read Full Bio Coverage Areas: ← Facebook Enlists Friend-Sourcing in Preemptive Strike Against Google […]

Comments are closed

Comments or opinions expressed on this blog are those of the individual contributors only, and do not necessarily represent the views of Gartner, Inc. or its management. Readers may copy and redistribute blog postings on other blogs, or otherwise for private, non-commercial or journalistic purposes, with attribution to Gartner. This content may not be used for any other purposes in any other formats or media. The content on this blog is provided on an "as-is" basis. Gartner shall not be liable for any damages whatsoever arising out of the content or use of this blog.