I recently discovered Klout and have had fun playing around with it for the past few weeks. Klout is a service that hooks into Twitter and a handful of other social networks and issues you a score from 1 - 100 that measures your online influence. If you are curious about the details of how they calculate their score, or how they define "online influence," you can read about it here. In addition to generating an overall score, they also measure your style, list the people you influence, and list topics that you are influential about.

Two weeks ago I had three topics on Klout that I was influential about -- "developers," "hurricanes," and "college." I'll gladly accept the "developers" topic as I am a former software developer and now manage them. The other two confused me and made me skeptical about Klout's ability to semantically parse the text that appears in tweets or status updates. "College" clearly comes from the fact that I have tweeted about CollegeHumor (I work there). I'll post a video that I think will resonate with my followers, or re-tweet a significant announcement (like when we recently published a book). What's odd about seeing "college" as a topic is that I have never used that word without the word "humor" immediately following it as one word with no space. And as for "hurricanes," I have one single tweet that contains that word. This tweet was re-tweeted twice and garnered one reply on Twitter. I also used this as a status update on Facebook and got 7 likes and 6 comments. A fine level of engagement, but not nearly as big of a response as many of my other posts.

Since becoming influential about "hurricanes" was so simple and unintended, I figured that it would be easy for me to become influential about any topic I chose. I decided that I would become influential about "soup." I like soup as much as the next guy, but I like saying "soup" quite a bit more than the next guy. So I threw down the gauntlet and gave myself ten days to get this done. While Klout updates a user's score daily, they only update a user's list of topics weekly so I figured this was enough time to talk about soup and allow Klout to process and refresh.

Looking back, I have twelve tweets in this timeframe that contain the word "soup" - sometimes more than once. Almost all of them were re-tweeted by multiple people and I got a number of direct replies as well. Sometimes the #soup hashtag was used, sometimes just the word, and often both. On both Facebook and Twitter I posted pictures of the soups I was eating during the week, and I got a handful of likes and comments on Facebook. I even posted about soup to my rarely-used Tumblr (which is also hooked up to Klout) and then tweeted that post. I was feeling pretty solid about this plan considering I not only posted about soup more than developers, hurricanes, and college combined, but because I got more engagement around these posts than I usually do.

I logged back into Klout after my self-imposed ten day limit was up and was shocked. Not only was I not influential about "soup," but Klout now had me listed as influential about "Klout." I immediately looked back at my soup tweets and realized that many of them also contained the word "Klout," but certainly not all of them. Surely if I were influential about Klout, I would have figured out how to become influential about soup, or at least understood why I had failed to do so already. Putting this Catch-22 aside, I started to look into this issue more closely and found that many people are reporting the same types of inconsistencies. This thread on Quora is people discussing how Klout has listed their topics as things they barely talk about while ignoring their true specialties.

One of the things I started to look into was if a topic could be any word or if there was a predefined list of topics. Clearly they must employ some type of filter to prevent words like "the," "and," or "fuck" from registering as topics. I asked on their help forum if they use a blacklist or whitelist approach and in the case of the latter, if "soup" is even a possibility.

I still have hope that one day soon I will be influential about "soup." My Blogger account is also connected to Klout so it's possible that this post and any tweets I make about it will push me over the edge. One nice side affect of this effort is that I have been eating a lot more soup than usual. If you have questions about the soup at Hale and Hearty or The Lobster Place in Chelsea Market, I am a good source of knowledge and can help you choose which soup may be right for you. If you seek details about the inner workings of Klout, you may want to look elsewhere.

In the introduction of "You Are Not a Gadget," Jaron Lanier recounts a story where he spoke at a conference and encouraged everyone not to tweet or blog during his talk: "If something I said was memorable enough to be worthy of a tweet or blog post later on -- even if it was to register violent disagreement -- then that meant what I said would have had the time to be weighed, judged, and filtered by someone's brain. Instead of just being a passive relay for me, I went on, what was tweeted, blogged, or posted on a Facebook wall would then be you."

The people I follow on Twitter -- mostly those in digital media or technology -- use it primarily for professional self-marketing. Occasionally these tweets contain insightful comments or interesting opinions about a news item. Sometimes they act as teasers for longer form content written by the person tweeting. Most of these tweets are just links to someone else's news story or blog post with a few meaningless extra words appended. I'm sometimes guilty of this behavior, even though I have always been of the opinion that it's not a worthwhile activity. Occasionally I'm inspired to share something I feel strongly about, but usually I'm just falling victim to the standard groupthink that you have to be active on social media to represent yourself professionally. I'm not against professional self-marketing, I just don't believe that simply being a passive relay of information will improve your online reputation.

One former co-worker of mine has a Twitter feed that is a non-stop barrage of re-posts from Techcrunch, Mashable, SAI, etc... This person developed a reputation for zoning out in meetings and turning his attention to his iPhone or iPad where he read and reposted news stories. It was clear to all of his co-workers from the timestamp on his tweets that he was doing this during meetings and all throughout the workday. In his efforts to establish an online reputation within his industry, he seriously damaged his reputation at his own company. If he took one of those stories and gave it some time and critical thought, I have no doubt that he could have produced an excellent post about it. Ironically, he also could have spent hours of time during the workday doing this and no one would be the wiser.

Lanier's line about being a "passive relay" is a very concise way of representing a type of online social behavior that I always thought lacked real value. I'm very curious to see if this opinion is codified in the algorithms that measure online social influence. I recently signed up for Klout and PeerIndex, two products that measure online social influence and issue a score to their users. The true purpose behind these companies will become clear when each amasses a giant store of data that every brand marketer will pay out the nose to get their hands on. Both Klout and PeerIndex have been clear that the number of people you broadcast to is not relevant to your score, and that they have ways of measuring your true reach and influence. They don't reveal more details about this, but there is no doubt that they are betting their future business model on a deep and powerful algorithm that will yield valuable data.

As I spend more time looking at these products it will be interesting to see if the people who act as a passive relay are rewarded for their behavior, or if these algorithms will view this behavior as I do. If it's the former, I think we can all expect a lot more useless clutter in our feeds since adding game mechanics to a relatively mindless activity will only encourage more of that activity. It would be impressive if Klout and PeerIndex rewarded those who took the time and effort to weigh, judge, and filter a news story and produced something that was truly them.