The Interest Graph: Smart Filters for Social Media

Social networks are undoubtedly changing the way we discover, consume, and
distribute content on the Web. And the proliferation of newsreader apps like
Flipboard and Pulse has made information more accessible – and consumable –
to a huge audience that never even heard of RSS. But although it’s easy today to
follow people and to follow publishers, we’re still failing to deliver on a key promise
of digital media: the ability to follow topics. The social graph is well served, but the
interest graph isn’t.

I’m a huge Facebook fan. My Facebook feed accounts for at least half of the news I
read and watch every day. It’s where I first learned about the Egyptian revolution,
the final score for the FIFA Women’s World Cup match, and Amy Winehouse’s
untimely death. And I’m not alone: a recent Pew Research study showed that
Facebook was the second or third biggest source of traffic to the top five nationally
recognized news sites.

Probably the best thing about reading news from your social feed (besides the
absolute zero set up cost) is that it’s chock full of interesting things you’d never
think to search for on your own. Sure, there’s a bit of overhead to pay in the form of
Farmville notifications and links to cat videos, but how else am I going to find out
about a road-legal, turbine-powered batmobile? That’s some powerful serendipity.

The downside is that you won’t often get updates on things that your friends care
less about than you do. Most of my friends aren’t gamers, for example, so I can’t rely
on Facebook to give me the latest news (or any news, really) from E3 or the GDC.

The current workarounds to this problem are generally variations on the theme
of mapping your interests to a set of publishers or individuals that frequently
post on those topics. Both Facebook and Twitter let you follow specific brands or
organizations (like Rat City Roller Girls or American Diabetes Association) to get
content updates pushed into your feed. So you can stay on top of your favorite
team, band, or company… provided someone has stepped up to be its official
editorial “sponsor” and curator.

On Twitter, you can even roll these accounts together into lists based on topic
area. This is exactly how Flipboard’s pre-defined channels work, for example:
FlipTech is an aggregation of posts from a select series of publications (TechCrunch,
ReadWriteWeb, AllThingsD, etc.) that focus on the tech sector. FlipGaming pulls
together everything from Gamespot, Kotaku, GamePro, etc.

This source-based aggregation is a good strategy for getting general-purpose
coverage of a broad topic area. But if your interests are more specific, e.g. “Cloud-
based computing,” not just everything “Technology,” it’s really difficult to fine-tune
a feed to bring you only the right content. You end up either with nothing at all to

read, or skimming through dozens of articles in a noisy feed to find the one or two
you care about. Robert Scoble recently called this out as a problem for the fledgling
Google+
, but ultimately it’s not problem with any one social network. It’s a problem
with the source-based approach in general.

Everything we do at Evri is aimed at solving this problem in a much more intuitive,
immediate way. We develop products that leverage your interest graph to discover
new content with the same ease and flexibility that your social graph provides
today. Instead of putting the burden on you to figure out who the authorities are for
each of your interests, you just tell us what you want to hear about. We listen to all
the conversations, read all the blog posts, scan all the tweets, and ultimately figure
out the most important news to show you for any topic.

And when I say “we,” I mean “our robot minions.” They apply sophisticated natural
language processing techniques to “read” every article and post and figure out
what it’s inherently about: who’s doing what to whom, where, and when? That
information is combined with other clues (like trends in overall popularity and
sharing behavior) to figure out what constitutes the “top news” for over two million
different topic areas.

Here’s an example: a trend that fascinates me these days is the rise of virtual
economies in social gaming. Companies like Zynga are making a fortune by selling
individual enhancements to online entertainment experiences – things like a bigger
engine or a custom paint job for a video-game race car. And as interested as I am
in this from the gamer side of the fence, I’m even more interested in the business
questions: who’s really capitalizing on this trend? How much are they making? What
sells, and what doesn’t? How does all this impact the real economy?

There’s absolutely no way for me to pull together an ongoing feed of this kind of
information from a publisher list. Gaming mags and websites may occasionally
report on overall growth or subscriber figures for a particular game, but it’s a very
small fraction of their overall content. Publications like Forbes & VentureBeat offer
some really insightful coverage, but again, this isn’t the only thing they’re reporting
on and I don’t want to skim through dozens of articles about the rise of mobile
computing or cloud networks.

Instead I subscribe to a channel on Evri: Social Games are Big Business. It combines
a bunch of abstract concepts (like “virtual currency”) together with some specific
company names to generate a feed that highlights the real-world bottom line
for this area. Things like mergers and acquisitions, analyst forecasts, earnings
announcements, etc.

That’s my personal favorite, but it’s far from the only example of how Evri hones in

on news by topic rather than by publisher. Check out Reality TV, Alternative Energy,
Venture Capital, or Celebrity Divorces. They all do a great job of filtering down a
flood of content from thousands of sources into an interesting, relevant stream.

As cool as it is to see these topical channels working in the context of a dedicated
WWW experience, I’m even more excited about bringing them to the iPad. The iPad
is such a great fit for this kind of content experience, and it’s really raised the bar for
how naturally we interact with the news we’re reading.

Scrolling through Web pages on your desktop got you feeling so Y2K? Imagine how
much cooler it’s going to be to lean back with your iPad and customize your own
digital magazine
, blending the great serendipitous content from your social network
with topical content that directly targets your interests. (How much cooler? In the
words of Andre 3000: “Ice Cold!”)

We’ll be talking a lot more about our iPad app launch in the coming weeks, so stay
tuned…

If you enjoyed this post, please consider to leave a comment or subscribe to the feed and get future articles delivered to your feed reader.

One Comment

Leave Comment