[Corrected 11/14/13, see below] In the world of news-reader apps, the individual reigns supreme. These days, nobody dares to release a news aggregation product for smartphones, tablets, or the Web without including some sort of personalization feature to weed out unwanted content. (The idea that editors should tell you what they think you need to know, unmediated by high-tech filters attuned to your preferences, is just so 20th-century.)
But what news-app startups mean by “personalization” can vary quite a bit. In Flipboard’s tablet and smartphone app, for example, you can sign up to see content being shared by the people you’ve friended or followed on Facebook, Twitter, and Instagram, or subscribe to “magazines” curated by users you trust. So Flipboard doesn’t know what you like, exactly, but it knows what your friends are sharing, and it assumes you’ll like the same things. A competing personal-magazine app from Zite, the San Francisco-based subsidiary of CNN, also taps the social graph, but it’s better known for learning your actual preferences over time, by asking you to pick topics you like and give a thumbs-up or thumbs-down to each story it recommends.
Nate Weiner thinks both approaches have drawbacks—and he’s got an idea about how to make personalization work better.
Weiner, pictured above, is the founder and CEO of Pocket, the San Francisco-based startup that operates a “read later” service used by roughly 10 million netizens. (In fact, the app used to be called Read It Later.) Every day, Pocket users save 1.5 million pieces of content that they find on the Web or in other apps—mostly news articles, but also things like images, videos, recipes, and product pages from shopping sites. They can then view these items at their leisure using Pocket’s apps, which are available on nearly every brand of tablet, smartphone, Web browser, and e-book device.
At a press event in San Francisco yesterday, Weiner unveiled version 5.0 of the Pocket app, which is available now for iOS devices and will be coming to Android devices next week. The upgraded app includes some nifty new features that help power users of Pocket sort through the (typically very long) lists of items they’ve accumulated and decide which ones to read.
But the big story of the day, at least in my eyes, was Weiner’s second bit of news, about a new program called Pocket Preferences. It’s a way for app developers outside Pocket to tap into what Pocket knows about the content preferences of its users, so they can do a better job of presenting personalized content in their own apps. In other words, Pocket Preferences is Pocket’s shot at owning the “interest graph,” at least as it applies to online content.
A takeoff on the idea of the social graph popularized by Facebook, an interest graph is a map of the topics important to a given individual; such graphs are at the core of many tailored services, such as Google’s personalized search results.
Already, the makers of at least two news-readers—Zite and Hack Later—are incorporating Pocket’s version of the interest graph into their apps. But what makes Pocket think that its Preferences feature will be attractive to even more companies?
It’s pretty simple, in Weiner’s eyes. Everything in your Pocket list is there because you saved it. That’s the purest possible signal that it’s interesting to you.
So, no more extrapolating users’ interests based on the interests of their friends. All Pocket has to do is analyze the items people are saving and extract topics and key concepts, and it’s got a thorough and up-to-date profile of the subjects relevant to each user. It can then export those profiles to other companies, which can use it to jumpstart, refine, or even replace their existing personalization algorithms.
After the press event, I asked Weiner whether he sees Pocket Preferences as a message to companies like Facebook, Twitter, and Flipboard that they’ve got the interest graph all wrong.
“It’s not that they’re doing it wrong, it’s that they don’t even have access to it,” he replied. “If you think about your Twitter feed, not everything that goes through there is something you are interested in. Just because your friends like something doesn’t mean you like it.”
Even apps like Zite that have their own ways of measuring user interest can benefit from Pocket’s view, Weiner says. “The problem is, Zite doesn’t know what you are interested in until you use it—you have to thumbs-up and thumbs-down, like on Pandora. With Pocket, you don’t have to think about that. Literally using it is what’s powering it. It ends up being a much stronger signal.”
To Weiner, Pocket’s other big advantage is that people use it to save a huge variety of items discovered across their digital networks. News articles make up only 55 percent of what users save. The other 45 percent is comprised of
