In the year and a half since Klout launched at SxSW 2009, the company has grown to become the leading social network influencer measurement within the online communications space. With success has come criticism within the public relations and advertising sector. This criticism of the algorithm-driven influence measurement ranges from its Twitter-heavy measurement of people’s participation to the ethics of its Perks program.
Klout CEO and co-founder Joe Fernandez conducted an open interview on Tuesday that included a dozen communications bloggers and influencers to discuss these criticisms. The call revealed a corporate culture that actively weighs and debates the merits and ethics of influence. Here are four ways Klout is looking at its challenges and how it may evolve:
Counterbalancing the Weight of Individual Data Points
When Klout launched it used Twitter data only, creating a metric that was over-reliant on one network. In recent months the service has added LinkedIn and Facebook analysis to its analysis of an individual’s influence capability.
Guest author Geoff Livingston is the co-founder of Zoetica, a social enterprise that provides superior communication consulting, training, and strategy to help mindful organizations affect social change. He has worked as a public relations strategist in the Washington, D.C. region for more than 16 years, and is the author of “Now is Gone“. He can be found at geofflivingston.com and @geoffliving.
Fernandez outlined the future algorithm roadmap, which includes adding the ability to weigh up to 20 social networks and commenting services to one’s Klout score. In addition to diffusing the weight of an individual data point with multiple networks, Klout may directly integrate individual content creators and systems. According to Fernandez, it is rare for a person to achieve a Klout score higher than 60 without creating some sort of content outside of just Twitter or Facebook.
“We do plan on tying in with the content systems,” said Fernandez. “We want to move beyond just ReadWriteWeb or TechCrunch, and focus on individual writers. Conglomerate blogs are pools of voices, they are media properties. How do we break the author’s influence away from masthead?”
By weighing a diverse set of data points, Klout is likely to bulwark itself against criticism about the algorithm’s validity. Past measures like Technorati and Twitter-specific solutions have suffered because of their inability to weigh multiple sets of data. Other measures – particularly those of blogs – that use multiple weights, such as PostRank and indexes like the AdAge 150, have survived criticism because they weigh diverse data points to counterbalance inconsistencies.
Build Volatility Into the Metric
Another strong criticism of Klout remains the appearance of influential relationships in one’s profile that may or may not be relevant anymore. Klout currently values long-term relationships over zeitgeist events, such as someone who garners sudden influence due to a particular issue or event. In the past, networks like Technorati and Digg have tinkered with their algorithm to de-weight long-term relationships and influence, which in turn have met significant criticism and even lower usage.
Klout wants to infuse more volatility into the algorithm to account for changes. The start-up is actively looking at how social data is created with more ease each day, and how data analysis can scale with it. Part of the timeliness challenge facing Klout is keeping the system fun and interactive for people to want to increase their Klout score while preventing gaming of the algorithm, which both Digg and Google search have had to combat.
“It’s important to build the algorithm so that it can’t be gamed,” said Fernandez. “It has to be volatile enough that it can’t be moved. You want the score to be responsive to real life, but you want to avoid intentional gaming like retweet networks. We have scientists that spend the whole day thinking about all of the loopholes. We are going to have to live with (a constant battle against) the overall gaming.”
“Influence is such a soft word,” he continued. “The idea of influence is everyone’s individual lens, and we are trying to standardize it. […] We see the idea of influence as the ability to drive action, and action can include retweets, mentions, and the ability to get people to do things like participate on calls and events.”
The encouraging aspect of the algorithm volatility issue is Klout’s adherence to the past. It does not want to repeat the mistakes of past algorithms, and actively looks at what has worked, too. Only time will tell if the company is successful, but it’s not operating in a vacuum.
The Warren Buffet Problem
If one were to type Warren Buffet’s profile into Klout, there would be no measure for him. Inside Klout, executives and staff actively debate the Warren Buffet problem. How do you account for someone who has real-world fame but no online influence?
This inability to qualify influence is endemic in any quantitative metic with limited data sets. Not every non-present celebrity has the presence of mind to create an account for references, like Seth Godin did. Further, there are individuals who may be influential in a specific sector, but not present in social networks.
But are these people truly influential online? Is the Washington Post influential on Twitter even though it gets retweeted all the time?
“Participation is an element,” said Fernandez. “And you look at the accounts retweeting the Washington Post and Mashable and there are a lot of bots and not so many influential people retweeting it, but the scale and speed that it occurs, you can’t deny there’s some influence.
“Influence is such a soft word,” he continued. “The idea of influence is everyone’s individual lens, and we are trying to standardize it. We struggle with this, and everyone’s conversation demonstrates it is a lightening rod. We see the idea of influence as the ability to drive action, and action can include retweets, mentions, and the ability to get people to do things like participate on calls and events.”
One of the ways Klout counters the qualitative issue is via subject matter-specific analysis of influence. Further, the company has a taxonomy of influence, thought leader, curators, etc. which allows it to parse what kind of a role they play in social networks. Most of these deeper qualifications and analysis are available online already.
The Ethics of Klout Perks
Klout Perks such as the Virgin America program are not only a way for companies to leverage the database and actively engage influencers, but also a method for Klout to test the validity of the algorithm. The company has built an ethics program around Perks based on several conversations with organizations like the Interactive Advertising Bureau and examining the FCC’s disclosure policies.
However, the taking of Perks has raised cries of foul play, some a matter of jealousy and other with a little more substance. In particular the issue centers around social media consultants who are responsible for recommending (or not) Klout to clients, but then participate in the Perks program.
“We’re not sending Scoble or big bloggers perks,” said Fernandez. “The way we’ve structured them is to make sure that people don’t pollute their Twitter stream. That’s the last thing we would want. But we’re in a space where everyone is a mid-tier influencer who is probably counseling someone. The percentage of end-users who are participants in Perks, and then use our disclosure recommendations is not as high as we’d like it to be.”
Like other matters, ethics remains a matter of debate and evolution at Klout. It’s clear that the organization wants to influence more ethical disclosure of its Perks program and is also examining its program on a consistent basis. The company says it values its place in the industry as an influence standard above the Perks program.
