Twitter "Klout": A New York City Election Case Study

Building on Nancy's post Wednesday about the interesting relationship (correlation is probably overstating the case) between the number of social media followers (Twitter + Facebook) tallied by some of the candidates running for various New York City offices and their showing in Tuesday's primary, I thought it might be interesting to ask Joe Fernandez, the founder of Klout, a start-up that analyzes people's Twitter profiles, to take a look at the two candidates in the run-off for the Public Advocate's office, Mark Green and Bill DeBlasio, and tell us who might be making better use of the platform.

As Nancy noted, DeBlasio edged past Green Tuesday, surprising many observers who expected the better-known Green (who was Public Advocate a decade ago, and has a fairly high media profile) to convert his higher name recognition into more votes. DeBlasio had almost eight times as many social media followers than Green, 3,265 to 445, before the primary. He also had the endorsement of the New York Times, and the grass-roots field savvy of the Working Families Party, a local powerhouse, also helping him turn out the vote. I'm not suggesting that DeBlasio's use of social media played into this upset (more likely it was a hidden indicator of his actual level of support from hard-core political activists, vs people responding to a public opinion poll). But it's interesting to take a look at Fernandez's analysis of their comparative standing and use of Twitter:

Looking at the data @Green4NY is using Twitter to simply broadcast his message (no different than any form of advertising really). @billdeblasio is doing somewhat better (but not great) at getting people to engage with him through @ messages and rt’s.

If you take a look at the content that @Green4NY has tweeted, in terms of information it provides it is probably better than @billdeblasio but the way he is delivering it doesn’t seem to be effective. There is not a lot there to connect with and I think he is missing the point/opportunity of twitter.

A glance at both candidates' recent tweets shows that Klout's algorithms are pretty smart. Green's recent tweets are basically all about his 100 New Ideas for New York. Each tweet is a pointer to an idea. While this is an impressive and creative list, it isn't very interactive or conversational. Whereas DeBlasio's recent tweets are pretty much in the candidate's voice, though in truth the typical tweet is something like "Visiting some wonderful folks at Elmhurst-Jackson Heights SeniorCenter." Not a lot to connect with, as Fernandez says.

I asked Fernandez if it was helpful for a political candidate to try to get followers on Twitter who themselves have a lot of followers. For example, would it be valuable to have a celebrity like Ashton Kutcher (who has more than 3 million followers) following that candidate, in terms of measuring his "klout"? His reply:

We don’t care how many followers each person’s followers have. Follower count is actually, we’ve found, one of the least indicative factors in influence. It’s definitely part of the whole puzzle but there are lots of other things we think are more important. We do look at how influential (with follower count being a part of that influence calculation) the people are who @ message and rt a person when calculating their score.

Fernandez is right to take this approach. The other day I was talking with Jim Gilliam, the wizard behind act.ly and TweetProgress. He told me that one day he got worried when he saw a Twitter celebrity with something like 800,000 followers using act.ly to push a petition, fearing that a surge in traffic would swamp his server. But no such surge happened. Jim's conclusion: most of those 800,000 followers are people who signed up for Twitter, clicked on some recommended users' names or celebrities they recognized, followed them, and then never used the site again. Fernandez has a perceptive comment about this phenomenon:

We actually calculate something we call “True Reach”. Personally I have almost 500 followers. Some % of those people are spammers or have inactive accounts. Another segment follows so many other people that there is no way my tweets are going to have any impact. Then there is a small amount of people who actually interact with me, tweet about the same topics and have overlapping networks that my tweets actually have an impact on. For me the count is actually 87 and when I think of the number of people that probably read my tweets that feels a lot more accurate than 500.

What can you learn from this? My biggest takeaway is that quality beats quantity in online engagement. Connecting to a smaller number of people in a deeper way is more valuable than simply amassing a big number of friends or followers. Participating in conversations, saying things that are interesting enough to be retweeted, reciprocating, getting retweeted by a growing group rather than the same few people, replying to people--these are all steps would-be influencers can take in growing their base on a social network like Twitter.