Is Yahoo building an 'interest graph'?

One writer thinks so. A look at how it impacts the tech giant's bottom line.
Written by Andrew Nusca, Contributor

If you haven't noticed, the Marissa Mayer-led Yahoo has been on a bit of an acquisition roll lately, buying up a number of small start-ups as it drives toward relevance. At least, that's the plan.

The latest? Summly, a mobile news reader startup helmed by someone young enough to not know a world that Yahoo did not inhabit. Sobering.

(The others: mobile interest tracker Stamped, news clipper Snip.it, restaurant recommendation app Alike, culture recommendation app Jybe and video chat platform OnTheAir.)

What's the bigger strategy, exactly? Writing for Fast Company, Sarah Kessler takes a stab at it. Her conclusion? Mobile, sure. But it's really all about building an "interest graph" -- loosely defined as a map of all the people, places and things its various customers are interested in.

(Contrast that with a "social graph," which maps the actual relationships people foster between them. In other words, Facebook -- at least, the friends part of it.)

She writes:

In order to paint your picture based on your interests, Yahoo needs to get to know you first. “When somebody invests time with you and your products,” [mobile chief Adam] Cahan explains, “they’re actually creating and changing the voice of the products back to them. They’re identifying the things they like.” [...]

It’s a loop: Increasing engagement increases personalization. Increasing personalization increases engagement. But to pull it off, Yahoo needs to focus on defining interests and providing users content that fits theirs perfectly--and that’s the thread that runs through its acquisitions.

An interesting take.

The magic word Kessler doesn't mention in her post? Advertising. In public statements, Yahoo has spun its latest tale as one about customer personalization; the truth is, this has been a pillar of its not-so-successful strategy all along, starting with search.

Here's former CEO Carol Bartz in 2009:

"Relevance is really our secret sauce… to what’s happening with people and things that matter to you most. That means we’re either letting you personalize what you see on Yahoo! (although we know no one wants to work too hard at doing that these days) or using what we know about to surface things that match what you care about."

And here's Search Engine Roundtable's Barry Schwartz paraphrasing a Jerry Yang keynote speech back in 2005:

Jerry Yang said that the relevance of search is the main component of any search. He said its not just about algorithms anymore, like PageRank or Link Data. Its more now about personalization and tying into that preference based searching. This adds a tremendous amount of value. My Yahoo! is a piece of this but you will be seeing it more.

So, personalization? Not new for Yahoo, and certainly not a new business concept. There's a reason you like it when your local shopkeeper greets you by name, and bristle when your local Starbucks barista requests it so that she can write on your cup...every week.

In its earliest days, personalization was a way for Yahoo to provide a better, more differentiated (search) product. Later, it was a way to secure all the customers it was attracting. In the middle years, personalization was bandied about as a way to save the company from the cruft, if I may, that had built up around its cluttered portfolio of consumer products.

Today, as it always has been, personalization is necessary for Yahoo to continue to conduct business. As of its most recent fiscal earnings report -- summarizing results for the full year 2012 -- about 43 percent of Yahoo's revenues are tied up in display advertising; another 38 percent is tied up in search. When you're making money from 1.) businesses who want to target specific people, and 2.) people who want to find relevant resources using your directory, "personalization" is pretty much the fundamental connective tissue. It is what roads and bridges are to a cross-country shipping company.

What Marissa Mayer is doing here, then, is following her customers. There are two major external trends that have negatively impacted Yahoo in recent years: first, the steady move to mobile computing, where monetization (and product development) hasn't been as lucrative or pressing; second, the complete overhaul of the way people navigate the Internet. Search was popular when the average Internet user was anonymous. Now, we all fill out individual user profiles with personal information so that we are served more relevant information from the get-go.

Yahoo is still catching up to all this. Despite its enormous digital footprint -- its homepage is still a top destination in the U.S., and it has lots of users in its database -- it has done a poor job of organizing that attention into groups of like-minded individuals it can easily sell against. And there's not a lot of loyalty there, either.

For example, a casual visit to the Yahoo Sports homepage right now shows an immersive advertising environment -- leaderboard, skin, footer, MPUs, for those of you who are familiar with industry lingo -- for Leinenkugel beer. There are also tiny ads for Netflix and State Farm auto insurance. While I do enjoy a Leinenkugel brew from time to time, here's the problem: it's a Wisconsin-based brewery that distributes heavily in the American Midwest. Its wares, therefore, are fairly hard to find here in Philadelphia. Further, I'm already a Netflix customer. And I don't own a car to insure. So, despite my willingness to provide ample personal information about my preferences, nevermind my location-specific IP address, Yahoo is serving me mostly irrelevant ads. Oh, and not a single headline on this webpage has to do with a team in my area.

It is clear, then, that Yahoo needs a means to obtain this personal information -- as well as reflect it in its products -- in a fashion that doesn't make it seem like it's building the next Facebook. (That ship has sailed.) This kind of infrastructure is critical for the company's future fortunes. At a time when user targeting is finally beginning to live up to the hype, it's not optional anymore.

Editorial standards