What If You Were Paid Based on Your Klout Score?


At GigaOM’s Net:Work conference on Thursday in San Francisco, Salesforce.com (s crm) chairman and CEO Marc Benioff said the company has an analytical tool that tracks activity on Chatter, its internal Twitter-style corporate social network (which the software provider also sells to other companies), and that Salesforce is giving employees who provide valuable information or contributions to the corporation via the network extra compensation, in the same way executives are given bonuses for exceeding sales targets. But is that really such a good idea?

“You, as an individual contributor, can receive the same compensation — stock, cash — as an executive vice president,” Benioff told Forbes editor Victoria Barret in an onstage interview. The Salesforce CEO didn’t provide any details about how the analytics tool (called Chatterlytics) works, or how the compensation for the Chatterati — which is what the most active Chatter users are called — is determined. But he made it clear that activity on the social network is seen as a crucial part of working at the company. So will other companies adopt this kind of model for their own internal compensation? And if they do, is that necessarily a good thing?

As social networking tools and features have been making their way from the consumer world into the corporate world, the pressure to somehow quantify that activity — to show that it actually has an effect on business metrics such as sales, profits and so on — continues to increase. Companies such as Klout and Peerindex have been trying to solve that problem by looking at Twitter and Facebook and ranking users based on who follows (or friends) them, who re-tweets them or re-shares their status updates, and tries to compile an overall influence score based on metrics such as “reach” and “amplification.”

Whenever there is a certain volume of human activity, someone is bound to try and measure it, so it’s not surprising that analytics would be applied to social networks like Twitter or Chatter. But it’s not clear (at least not to me) that it’s all that easy to determine the value of that activity in business terms, not to mention how (or who) to compensate as a result of that activity. Should it be the person who spends the most time on the network? Should it be whoever posts the most? In some cases, a person who posts one or two messages might actually provide more tangible value — but how do you measure that?.

As more than one person has noted, including Chuck Hemann of media and communications company WCG in a recent blog post, your Klout score — or your Chatterati score, for that matter — probably isn’t a great measure of your overall influence anywhere other than Twitter, which only a fairly small proportion of people use, as a Pew report this week pointed out. That’s not to say that Chatter or Twitter or any other social network isn’t going to help employees collaborate inside companies like Salesforce and elsewhere — but it’s another thing to say that the people who spend time on those networks deserve to be compensated differently than those who don’t.

If Salesforce is using Chatter activity as just one of a number of things it looks at performance-wise when it determines how to reward employees, then that seems fairly reasonable. But making social networking alone a proxy for how much value someone adds to the company seems to be going down a potentially dangerous road, with unpredictable consequences. Wall Street bonuses certainly triggered some unpleasant behavior by traders and bankers during the recent financial meltdown — companies need to be careful about what kind of behavior they are motivating when they think about compensation.

Related GigaOM Pro content (sub req’d):

Post and thumbnails courtesy of Flickr user Andres Rodriguez


Salim Virani

The world is full of examples where compensation is biased towards the measurable and demonstrable. The measurable and visible parts offer the most practical way to improve overall. This isn’t something specific to social networking.

I’d suggest it’s erroneous to dismiss a measurable subset of activity on the basis that it doesn’t necessarily represent the whole. If improvement is your goal, it makes sense to focus on areas that are measurable, even though the non-measurable areas might not be improved.

For example, your website conversion rate doesn’t represent what happens offline, but since it’s more easily measured, it’s easier to improve than a offline marketing. Just because your online customers don’t represent all your customers, we still measure, improve *and compensate* online. And we certainly don’t worry about this being unfair to, or negatively affecting the behaviour of offline customers. After a decade of Google Adwords, we’re comfortable with this concept – now we’re seeing the same opportunities in social networking.

Using Klout, you may only improve behaviour among the Twitteratti, but you are still improving behaviour. Whether or not this is fair, it works.

David Pylyp

I wish that I could be paid extra for my Klout score.

It would make credibility and brand management so much easier. But alas I am just a Realtor selling Houses and Condos in the GTA. Its not the google ranking but the influence I can generate.

All good points How can we be monetized?

David Pylyp
Living in Toronto


Can we have a followup interview, with a little more time spend and focus on how all of this fits together.

I for one don’t see it as such a big obstacle to determine the value of a given message if the context is known(which should be easier in a “closed” system). It should be similar to determine the information value for a given individual for any given article. But in a company the distributed nature of a given strategy/memory has to be factored in, which might be a challenge. Do they even have a context identifying system?

DARPA wants to identify soldiers before they become suicidal which most people also think is impossible. Let’s see how far that can be pushed, should be a similar complex system.

One other problem might be data organization, how does a “flat” organization cope with hierarchical data structures,conventional HR, sales, strategy and …. with the supporting applications also build on these assumptions.

Fun: More questions than answers, let the conversation begin.

Jeff Ross

I’m all for allowing some way for valuable ideas expressed through social media or any other means to result in added compensation or whatever is motivating to employees. I wouldn’t necessarily limit opportunites for reward to only those participating in social networks, though. However, if that is the company’s chosen avenue for expressing such ideas and everyone knows it, then it’s fair. Participation in the network would surely be driven to some extent by the potential reward.

Chuck Hemann

Matthew – Thank you for providing the link to my post. As I noted there, and you are noting here, off-the-shelf influencer analytics tools are generally lacking in their ability to actually define influence. The broader concern I have, and it’s something you’ve raised here, is that we’re coming up with channel specific influencer models, which is not how the Web operates. I’m an active (not as active as I’d like to be) blogger, Twitter user, Facebook participant, etc… Just because I have a certain number of followers on Twitter doesn’t make me influential across the social web. It MIGHT make me influential on Twitter.

Furthermore, what these tools is brand context. I don’t care how sophisticated the algorithms are they can’t replace the knowledge a human brings about the brand. Said another way, relevance is key. Reach is important. Multi-channel participation is important. Relevance is important. All things, to date, these kinds of influencer “products” have failed to deliver on completely. Will we get to a place where they are better? Probably. I never doubt the ability of technology (and the people behind it) to advance. Are we there yet? Absolutely not.

Thanks again.


Mathew Ingram

Thanks for the comment, Chuck — those are great points, as were the ones you made in your post. I think you are right that we will probably get there eventually, but for now the metrics we are using only tell us a tiny part of the story about what influence really means and how it works.

Comments are closed.