Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
Anyone familiar with the concept of big data should know that social media are a great source of information on consumer sentiment. But the next wave of analytics in social media will be influence.
It’s easy to categorize friends and co-workers who are concerned with their Klout scores or number of Twitter followers as vain, but the truth is that this stuff matters. If you’re a marketer looking to get your message in front of the right people, or a politician trying to stoke the campaign fire, figuring out which voice will best help spread your message is a lot easier than trying to connect directly with the entire tweeting world.
I was reminded of the importance of social influence this morning when reading a post by Greg Finn on SearchEngineLand about the Microsoft(s msft) Bing search engine, and how its algorithms factor in users’ social media activity “to merge a little bit of the search and browsing intent into one.” What Bing is trying to do is figure out what users are interested in and, therefore, what they’re actually trying to find with their search. (This is something Google(s goog) is trying to do as well, despite surveys showing people really don’t like the idea of personalized search.)
More interesting on the analytics front is how Microsoft determines how social media will influence search results. Facebook friends, of course, get special treatment (friends’ endorsements have been compared to celebrity endorsements in court), but Microsoft tries to determine author influence for Twitter using the following factors (quoting Finn):
- Ratio of followers to following
- Sharing (as a strong[er] indicator than just a like)
- Having good followers
- Following good accounts
- Tweeting about relevant topics
- Authority and Relevance of Retweeters
Some of these factors no doubt go into the Klout algorithm as well. There, business partners pay Klout to put their “perks” in front of the right set of people. The theory is that these influencers will try the products or services and, if they like them, express their feelings via social media.
I assume many companies are already taking similar approaches to using Twitter as a marketing campaign. Step 1 might be finding out how people feel about a particular product, show, etc., by analyzing the Twitter firehose. But Step 2 should be finding out which Twitter users are influential in that space and trying to make them happy. Or maybe part of Step 1 is weighting sentiment based on who expressed it — an influential voice coming out in support of or against something might be worth more than someone with relatively low influence in determining how something will play out.
But social media influence has import beyond just selling stuff: It also could help decide elections. It’s no secret that presidential campaigns are deeply engaged in social media and are mining the data those efforts generate. And, as my colleague Stacey Higginbotham recently reported, the Obama campaign is already using tools like Hadoop to analyze unstructured-data sources that include social media. As their methods get more sophisticated, it seems very likely that campaigns could start trying to identify and target social media influencers as well.
The whole tech world doesn’t need to follow you, support you and retweet your tech policies if @Scobleizer does.
As just about every discussion at our Structure:Data conference next week will address, big data ultimately boils down to efficiency. Companies and organizations want to learn as much as they can so they can get the most bang for their buck. There’s nothing more efficient than letting someone else do the work for you.
Image courtesy of Ronald S. Woan.