Monetizing social media means navigating ‘big, sucky data’

Solariat Founder and CEO Jeffrey Davitz has a message for anyone trying to leverage social network data to make money: “The fundamental problem with social is yes, it’s big data, but it’s mostly big, sucky data.”

What he means, he explained during a recent interview, is that life isn’t too easy for social platforms such as Twitter and Facebook, where deciphering what users actually want means poring through a lot of extraneous information. Google is able to earn so much advertising revenue because its search users are expressly seeking information on a specific topic. They’re relatively likely to click on a sponsored link if it will answer their questions or connect them with the products they’re seeking.

The challenge on those other platforms — for both the platform providers trying to create targeted advertising models, and brands trying to engage with consumers on the platforms — is getting messages in front of users in a manner that’s “congruent” with what they’re already doing. Take Facebook and its advertising revenue woes, for example. If users are generally on Facebook with the intention of interacting with their peers, they might not notice or care about the display ads lining the page, no matter how much data they share via profiles, posts, Likes and other interactions with the platform.

Engage when and how consumers expect it

During a panel at last week’s Data Science Summit (in which Davitz also participated), Dan Neely from Networked Insights described this challenge as figuring out “how to he part of the distraction.” If a company is just guessing when and how to approach customers, the company is just another entity — along with instant messages, wall posts, other brands, etc. — competing for that users’ attention. And if users are there to do social networking, corporate messages are probably going to lose that fight.

Davitz thinks there’s a way for social platforms to overcome this problem by using techniques such as natural-language processing and machine learning to identify those instances where users really are expressing “query-like intent.” It will never be as clear as entering “best hiking shoes” into a search engine, but, for example, someone certainly might note in a wall post or a tweet that he’s going hiking and needs new shoes. He might specifically ask friends which shoes they prefer. If you sell hiking shoes, there’s your signal. Rather than simply peppering someone’s page with ads about hiking because he listed it as an interest, now he’s actually in the market for gear and might pay attention.

Davitz claims proof that this approach works. His company, Solariat, is experimenting with a publishing industry partner to place content in front of Twitter users when they express interest in areas on which the publisher has an existing body of information (Solariat has done the same thing with companies that task a community manager with monitoring social media). Clickthrough rates are “astonishingly high,” he said — over 20 percent — and even users who don’t click aren’t marking the tweets as spam. Because the users getting replies are actively seeking information on a topic, even if they’ve only implicitly acknowledged as much, a pointer to that information is either welcome or, at least, seems natural.

Balancing effectiveness and creepiness

As with many big data efforts, though, the obvious concern with this approach is perceived privacy violations. “People might find it a little creepy if you had an overactive agent that was responding every time you say something,” Davitz acknowledged. In that case, companies might be better off gathering signals about users and targeting them with messages only at ideal times, which likely will vary based on the types of platforms and content involved.

For example, Davitz, an artificial intelligence expert, first got interested in social media engagement while working on a closed network within the U.S. military. In that case, personnel were ordered to sign up and had very low privacy expectations, which meant they expected to be monitored and regularly presented with additional information. That type of situation certainly had its advantages in the dynamic Iraq theatre of war, Davitz said, because soldiers posting about heading to a particular region, for example, would automatically receive the latest information on what was going on there and what to look out for.

Davitz thinks public platforms such as Facebook or Twitter, as well as ecosystem players such as Buddy Media and Vitrue, could learn a thing or two from how the military used social media as a channel for conveying information. If revenue is the goal, social networks have to be more than passive places where people just meet and interact relatively free from any corporate messaging, he said. They also need to look past ads as the only way of reaching consumers and perhaps find more natural and appealing ways to provide information.

A military-level of interaction might never be welcome on public platform, Davitz said, but “the thinking [on public networks] is really quite primitive compared to the thinking I saw in the military.”

Feature image courtesy of Shutterstock user FuzzBones.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

  • Connected world: the consumer technology revolution
  • Finding the Value in Social Media Data
  • 12 tech leaders’ resolutions for 2012



GigaOM