Gravity launches API, ushers in a personalized Web for all

By Michael Carney , written on February 1, 2013

From The News Desk

When new Yahoo CEO Marissa Mayer spoke at Davos of her plans to develop an “interest graph” at Yahoo, she described the content personalization technology as being “a few years away.” The group of LA entrepreneurs behind Gravity would take umbrage with that claim, as they’re already three years into building what many consider best-in-class interest graph technology and they've have been telling anyone who will listen.

For the last year, the company has worked with prominent publishers such as The Wall Street Journal, CNNMoney, and TechCrunch to deliver customized content in real-time to each of their readers based on their individual preferences. Gravity now delivers more than 25 million personalized content recommendations per day to more than 200 million Web users.

The results have been impressive, with partner sites seeing greater than 200 percent increases in click-through rates, loyalty as measured by monthly return rate up 300 percent, and pageviews increasing in some cases by as much as 40 percent. Today, the company is launching its Personalization API to make the technology available to all websites and applications.

“We hope this will be the end of the one size fits all Internet,” Gravity co-founder and CEO Amit Kapur says.

Gravity works by combining natural language processing and machine learning algorithms to interpret the meaning and connections between various pieces of online content and user behavior. The system gathers information about each user’s behavior across its network of partner sites through their content consumption history, comments, likes, shares, Tweets, recommendations, and that of their social graphs. Gravity continually calculates the varying strength of a user’s interest in a given topic over time and returns recommendations designed to optimize engagement and user experience.

The end result is Web experience custom tailored to the individual user level. My Wall Street Journal is not your Wall Street Journal, because we care about different things. Eventually, my may not be your

Unlike "customization" technology of the past, which requires users to manually select categories and subject matter that they're interested in before content can be tailored to their tastes, Gravity works implicitly, and in the background. It demands no change in user behavior. And if it does its job correctly, the user never notices it's there. Ideally, they just enjoy the Web more.

With today’s rollout of its Personalization API, Gravity hopes to usher in the era of mass personalization. As Mayer hinted, personalization will be central to the future of the Web. Currently, most offer up a crude approximation of a personalized experience based on the social graph. This implies that the preferences and habits of our friends are an accurate proxy for our own preferences and habits. Mayer’s Yahoo and Gravity both agree that the social graph will give way to the interest graph, making this personalization much more accurate, and thus more useful to consumers.

Now two plus years into gathering Interest Graph data, Gravity says its understanding of the online consumption behavior and preferences of individual users is better than ever. As new sites join its personalization platform, those sites will benefit from all that institutional knowledge.

In other words, a new sports publisher doesn’t implement Gravity and start with a blank slate on day one. It starts with knowledge about its readers’ consumption behavior on existing Gravity sites. The more publishers that join this network, the more powerful it becomes. Implementation is free for publishers and is supported by the company’s sponsored story advertising.

Gravity has raised $20.6 million across two rounds of venture financing, with the lastest, a $10.6 million Series B round, coming in October 2012. The company is backed by GRP Partners, Redpoint Ventures, and August Capital. The company, co-founded by Kapur and his fellow MySpace alumni Jim Benedetto and Steve Pearman, has grown to 30 employees, with offices in Santa Monica and New York City.

While it's most known for what it's done in media, a more interesting application for Gravity may be commerce. The problem with daily deals and flash sales space – aside from merchant fatigue and declining margins – is that consumers are overwhelmed with choice. We’ve all been sent an email offer for something that’s clearly for the opposite sex, across town, or simply something we'd never do. If Groupon, LivingSocial, Gilt, JustFab, Zappos, and the like implemented Gravity, online shopping and discovery would elevate to a whole new level of enjoyable. This is all the more important in an age of mobile commerce where a smaller screen means retailers have to serve up the most likely results first.

The biggest risk for Gravity has always been the “privacy issue.” While the company does not store any personally identifiable user data, and doesn’t sell one publisher’s data to another, both Web users and publishers tend to have initial reservations about the technology.

Many publishers are quick to come around when they see the performance data. Consumers on the other hand are less cut and dry. If you asked the average website visitor if they wanted a company to follow their every move online, draw conclusions from that behavior, and then decide what content they see and don’t see, the answer would be a predictably horrified, no!

But as we’ve seen countless times in technology, people really don’t know what they want until you give it to them. Judging by the increased click through rates, users prefer Gravity's results whether they like the way they got there are not.

Every site powered by Gravity offers the option to opt out of personalization, but consumers do so at a negligible rate, Kapur says. Whether they are aware that they have this option, or that they’re seeing personalized content to begin with, is another question entirely.

As a consumer, I’m perfectly willing to share my personal data, so long as I’m clear what will be done with it and the value I’m getting in exchange is compelling. This is why I use gMail and Facebook – along with more than a billion others on the planet.