Facebook versus Semantic Web

Architecting Better Data-driven Digital Experiences

Facebook versus Semantic Web

One Step Forward, One Step Back
From my standpoint as a marketing science data junkie, the Facebook changes announced at F8, while very cool from my perspective as a user, ultimately amounted to no change at all in my role as a data analyst. While personally I am already enjoying the new design and sharing elements, when I put on my marketing hat, I can only be disappointed that in their version of the “open graph”, Facebook remains the only party with full insight into any users’ integrated history, and the exchanges across the social graph created by its users.

What did not change with the redesign is that Facebook’s business is still built on targeted marketing. Its expansion of interest signaling from just “liking” to now any verb will certainly improve Facebook’s ability to target based on unique and shared interests. Every business on Facebook would benefit from an understanding of their consumers’ shared interests and key influences across their social graph, but Facebook retains a tight hold on their sole position as market-maker.

A Self-contained Commerce Engine
With their changes in user experience and interactive capabilities, Facebook is seeking to solidify its place as the one true personal portal. In this context, marketers need to immediately begin a shift from thinking of Facebook as a forum for messages from brands that users have liked, and/or a system for serving targeted ads. What Facebook ultimately wants to to with its data is drive highly targeted and personalized apps serving every sort of commerce, and I’m sure they have a plan to extract some value for their role in each exchange.

And while each app owner will have access to whatever consumer insights they can capture directly, Facebook will (as is their complete right) keep tight hold on the really valuable insights; the interactions and shared insights of users connected through the social graph. Facebook Insights offers a very shallow level of insight into users and their engagement in the social graph. The Facebook API and Facebook Query Language (FQL) have offered some opportunity to access more depth in users’ interests and interactions, but access to the true depth of insight into preferences, influences/influencers and social interactions available to Facebook itself has always been stifled by API limits and awkward FQL indexing schemas, and as of now, there is no new API/FQL documentation around the recent changes.

Your Data’s Hotel California
I completely understand that marketers have no right to the data that Facebook has collected, so there is no real basis for protest from this standpoint – we data-driven marketers must take whatever Facebook is willing to offer.

Facebook users, on the other hand, while also opting in to the network, should still be increasingly wary of every new offer from Facebook to hold more and more detailed information about their history, preferences and relationships. The exchange for the new and fun ways to express your “true self” and interact with “friends” is that this can only happen in Facebook. While it’s true that all of the information you enter into your profile can be exported (in Facebook’s token nod to calls for data portability), none of your history of activity on Facebook – your likes (and now other verbs), the comments you’ve made on other posts – none of that stuff that is the true core of your Facebook use and history, that is of true value for marketing insight, none of that can actually be exported. That history is not yours, unless you stay in Facebook. (To paraphrase the Eagles, you can check-in any time you like, but your data can never leave).

The Ideal: A Semantic Web & Personal Data Portability
In a true semantic web, each individual could access and share any part of their prior history of digital interaction as well as current interests and preferences within and across that web with any party or site, and for any purpose, be it social exchange, play, commerce, etc. Facebook is becoming a really compelling place to hang out online. Unfortunately, Facebook’s current model seems aimed at becoming the only place to hang out online, at least if you want a web that understands your preferences and networks.

While Facebook’s changes will allow me to use apps that deliver behavior and preference information INTO Facebook (i.e. to document that last Saturday Scot read the new Neal Stephenson book, listened to St. Vincent and ate at the Phoenix Dim Sum restaurant), it does not easily allow any of that data to come back OUT of Facebook for use by other apps (i.e. a dining app that can take my FB history of “ate at” and use that for suggestions, or can send the stuff I’ve marked “read” to Amazon.com for recommendations, or can send my “listened to” history to the online station of my choice for custom playlists). For now it appears that my history for each type of verb can be kept siloed in each of the apps I use for each verb (and I can potentially export or share from each app) but should a smart new app come along that could recommend activities for me around town based on three variables; my listening history, my reading history and my check-in history, while that combined behavior pattern will be recorded in Facebook, it won’t be accessible for me to share with this app outside of Facebook.

All Nodes Lead to Facebook
In ancient Rome, it was said that all roads led to Rome. Well, for now anyway, it appears that in Facebook’s view of the social graph, all nodes lead to Facebook. Marketers may wish that they knew everything that Facebook does about how people behave and interact within and across the network, but digitally savvy individuals should really be looking for ways to take back control of their digital history and presence, with decisions about how and where it is used being made by them.

David Segal has created a video to share his vision of the personal data locker, within which individuals could collect and manage any and all digital records related to their world. Everything composing our digital footprint, from music playlists and shopping wishlists to health, education and financial records, would all be stored in one place and managed by one party – ourselves.

Anyone who has read this far should take the next step and check out the Locker Project, the Data Portability Project and the schema.org projects. The internet is cool because its wide-open nature makes anything possible. Let’s keep it that way, while demanding that our history and digital persona should belongs to us.

If Facebook’s business plan does not involve serving as a hub for the exchange of information, but rather Facebook intends the be the primary terminus for all user information, then hopefully a next generation of apps will be built to pull and mash-up preference history, behavior data and social graph connections directly from the same apps on which Facebook is basing its efforts at hegemony.

Portions of this post will be included in a collection of views on the recent changes to Facebook posted on Critical Mass’ Experience Matters blog. I encourage you to check it out.

One Response

  1. Bruce van der Kooij says:

    Wow. Great stuff.

    “Digitally savvy individuals should really be looking for ways to take back control of their digital history and presence, with decisions about how and where it is used being made by them.”

    Yes! And as far as I can tell Google Plus isn’t any better. In fact I can’t even get Takeout to work.

    I guess this is why there was so much interest in Diaspora. But I find myself rather disappointed with what’s come out of that project so far. Apart from an activity stream https://joindiaspora.com/u/ilya I can’t even see what they’ve come up with so far.

    And I don’t really want to grab the code, follow the installation instructions etc. to find out.

    As for the semantic web.

    I think I first heard about the term “semantic web” when I was studying software engineering in college about 8 years ago. Today I still find myself rather confused as to why the semantic web as envisioned by Tim-Berners Lee hasn’t arrived.

    If I look around I still see a lot of data being produced that computers have a hard time understanding (this comment included). I think the major problem is the absence of easy to use tools integrated in existing workflows that people can use to give data meaning.

    Freebase for example is an amazing project that I believe could have (had?) so much more impact. The front-end used to describe topics according to the Freebase schema is quite user friendly.

    However, today, almost 2 years after Google acquired it, as an outsider it seems to me they haven’t done anything noteworthy with it, such as actually integrating it in Google Search.

    I want to see if I can help in realizing the dream that Tim-Berners Lee as early as 1998. But even as a programmer (though only a mere journeyman) I have a lot of trouble making sense of how to use various semantic technologies in my own applications.

    For the project I’m currently working, which is in part a (technological) clone of Freebase + extra tools to integrate it in existing workflows, I have a lot of trouble thinking up of the database schema.

    I’m also yet again “stuck” using a mere relational database, while I feel I should probably be using a RDF store (or triplestore, or graph database).

Comments are closed.