Why Everyone is Now Talking About Klout

And you still don't know what you don't know.

Ever since I published my post a short two weeks ago, and Joe Fernandez, CEO, Klout, was kind to discuss the company's and tool's genesis at lunch the day the post published, many have entered the fray. Geoff Livingston was writing about it on the very same day, too.

Of course, it is a bigger conversation than any one tool, and there's room for everybody to comment and add their experience. This not being a popularity contest, or a contest of any kind, big props to Geoff for staying ahead on the story and for linking to other resources besides this one.

Who doesn't love a good conversation? For this week's post, we'll get right to it with two of the most frequently asked questions.

Can influence be bought?

One fundamental tenet of influence -- the broader conversation we'll be having leading up to our face to face in March -- is that it touches those who do affect a desired change in behavior. There is no expiration date on that, although if you ask a marketer, it would be as soon as possible, this quarter, even.

While influence may hit apparently suddenly, enduring influence has a long tail -- before and after it hits, preferably long after. Let's suspend our focus on influence for a moment though so we can see the bigger picture. This is a conversation about people, isn't it?

In that case, to stay on the practical side of things, it has been demonstrated that people can be bought. It's even in the finest art of storytelling. There is a global and local tradition on that, and not just for monetary gain.

In fact, you don't even need to pay in many instances, as the human desire to compete can be triggered for free. You do know we test ideas and dynamics here. If influence can be bought, can it be sold? I'll leave you to wrestle with that one, while we look at how it can be gamed.

What do the tools measure?

There are more tools that purport to measure online influence in addition to Klout:

• PeerIndex -- described as a social capital system for measuring how influence is used, with a known bias towards technology and businesses. Of course, your mileage may vary depending upon being a registered user. From the description on their site, it sounds like if you have a breadth of interests and topics, you dilute your score. Two potential pluses, content farms and link blogs have lower value, and spammers are penalized. There is a bit of secret sauce about the algorithms, and the tool says it measures five areas: authority, resonance, audience, activity, and realness; which it then normalizes agaist the "top people" it tracks.

*Note: the tool seemed to have cached quite an old version of my account; because the topics I share are not in one of the buckets tracked by this tool, it categorized my activity as arts and entertainment, missing business strategy.

• Twitalyzer -- introduced as an analytics tool for comparing relationship strength. As of January 2011 Twitalyzer has become a largely month-to-month paid service. Business and agency accounts can track up to 100 streams. What the tools tracks are: impact score, influencer type, most commonly used hashtags, recent topics, and network on Twitter. 

*Note: when I looked at my own Twitter activity, it pulled from an old version of my profile, and showed people I never even @ replied as part of my network. It didn't pick up the #kaizenblog tag, even though I used it heavily for 18 months.

• Edelman's TweetLevel -- developed by the agency for its own research, it measures an individual's importance on Twitter. There are four result metrics: influence,  what you say is interesting and many people listen to it. This is the primary ranking metric; popularity,  how many people follow you; engagement,  you actively participate within your community; and trust, people believe what you say.

*Note: my handle gets high popularity and engagement scores, while trust levels could be upped. In other words, retweet more of my posts, will you?

• TwInfluence -- from the site, twInfluence is a simple tool for measuring the combined influence of twitterers and their followers, with a few social network statistics thrown in as bonus. It calculates two types of ranks: Reach Ranking, as compared to other accounts the site analyzed, and Relative Scores, where twinfluence calculates three derived statistics - velocity, social capital, and centralization - that don't really make sense without additional context.

*Note: the site didn't work at all for my handle. And I have a hunch that the explanation wouldn't have made sense anyway, given how the derivative statistics are described.

What does Klout measure? Using more than 35 variables, the secret sauce, on Facebook and Twitter, it measures True Reach, which excludes spam bots and inactive streams, Amplification Probability, which it says it's the likelihood that someone will act on your content, and Network Score, which is the influence level of your network.

*Note: last time I checked, it picked someone who influences me on Twitter I have not interacted with in months, if not years. My main activity on Facebook is focused on the blog's page, and not my personal profile.

How is Klout different? I asked Joe Fernandez to share his thought.

Honestly, I don't look at any of those tools as competitors to what we are trying to do. Our mission is to help individuals understand and leverage their influence. We are people focused. My belief is that if we can help consumers find value from their Klout score (whether it's through discovery, becoming more effective at social media or getting perks from brands) then brands will benefit also. Think of Facebook. The goal there was never to build a service for brands but to create something that people value.

We believe we can become the standard measurement of influence across the social web by going bottom up through the people.

I think the tools you listed are in for a rough go of it in 2011. TweetLevel is already basically abandoned (things like this inside big companies never work), Twitalyzer has a nice little dashboard and integrates our data but they are in a crowded market where it's a race price wise to the bottom. Twitalyzer is a side thing for the guys in that company so they'll be fine. PeerIndex are smart guys but I think they might have shown up too late. I think they'll find an interesting niche but it will require some pivoting.

A better question is what do these tools fail to measure?

In addition to what we discussed in our conversation a couple of weeks back, there is no reliable qualitative information, as well as definite correlation to a desired change in behavior and actions in the community, even as in some cases, they give the impression that something may be happening or has happened -- a retweet, a link shared, etc.

It seems to still be related to those other impressions that mean so little to connecting trust and authority with relevance, and ultimately action.

Why everyone is now talking about Klout

Because the tool hit a nerve early on, it got proper attention by becoming synomoum with the term online influence -- as uninformed as that may be, people are busy and like to simplify and respond to a company name like Klout.

It got Joe Fernandez and the team a nice boost in publicity along with giving them potentially enough incentive to clarify what the tool measures vs. what it does not.

There are plenty of people who have been doing amazing work on documenting how influence moves through relationships, and who have built influence algorithms out of passion and a desire to understand better how influence moves.

Klout scored (no pun intended) attention because it became the symbol for the whole conversation. A couple of good comments to Livingston's post highlight basic nature of what Klout measures and where in different ways:

In the end, unless the greater public might embrace this as an activity of merit to collect more and more people willing to click through on an empty action (because of a headline, linkbait, or fan of the subject matter but not the author), I suspect it will become an increasingly small pool of individuals who have a vested need to improve their scores. [Rich Becker]

I do rather think that the “influencer” model is still viable with some regard. But I’d be inclined to argue that it has to adapt to the new media. If Sarah Jessica Parker is one of the only people talking on television about hair color, that gives her 30 second spots a bit more impact. If Sarah Jessica Parker blogged about a hair color one time – well, let’s just say her SEO would need to rock for that to stand out in the ocean of online posts about not only that hair color, but the manufacturer, the problems, the customer service issues, and a myriad thousand other posts that will drown it out. No matter how many people read her blog? It’s still a drop in a bucket. But the same woman, same product, why not same level of influence? [Lucretia Pruitt]

Will Klout the brand suffer the long term repercussions of succeeding in being on everyone's mind so quickly, while trying to be all things influence to all people and organizations? Will you or your business suffer because you're focusing too much on one data point and how to game grow it?

Next week we'll talk about five influence traps you and your organization must avoid.

 

[image of book cover, Amazon affiliate link. Image of First Run. Image of Tree on Henry House Hill.]

If you enjoyed this post from Conversation Agent, subscribe, share and like it.

Comments

[Eric Pratum]

Now, folks from Klout might correct me here, but one of the things that makes Klout not valuable is the lack of transparency. What causes one number or another to shift up or down?

I just looked at my score. It looks like mine bumped up 1-2 points over the last few weeks and then dove 10 points suddenly just the other day. Is this due to something I've done? I would assume so, but honestly, I don't know why because I can't see an explanation anywhere.

[June]

I look at a majority of the aforementioned tools as a basic "grain of salt" measure. And influence is measured in the eye of the beholder.
Meaning what are you trying to influence? A purchase, action or what ever that may be. If you have a mass following and the ROI average is a low percentage yet your influence is high (rt's shares) but results in nothing other than consumption than are you influential compared to focused following with high ROI, less RT's and shares.

This said the tools can help to see if your on message & within your interest niche. In the end for accurate measurement other factors need to be within the tool set that are unique to the individual and their goals or seeking to accomplish.

[azeem]

Bah!
Our bad. We've updated your profile, it'll take 12 hours to work its way through the cache.

As for topics - those topics our estimates. We'll introduce some capabilities which will allow you to fine tune them in the near future.

And I think you've put your finger on something - which is the notion of influence is very watery. See what @joefernandez has to say on the use of the word here:http://twitter.com/#!/JoeFernandez/statuses/21996089720705024 in response to my point http://twitter.com/#!/azeem/statuses/21995511389102081

It is a first stab and it covers many different expectations and interpretations.

Perhaps the best term I've have seen is the use of 'eminence' which covers off many of the sources of eminence (from being a connector, being influential, having achieved, having expertise).

Our view is that people are building up social capital--that is a reputation based on different activities in different spheres--we currently measure your relative levels of capital on a series of online channels. The fact that you have built up this capital is indicative of something - and not something as hard as the fact that you might persuade other people to buy a product. But is a a measure of the extent to which other people are aware of you, would recognise you, might act on what you see, might believe what you believe.

The next stages in this evolution is to make this data more predictive of certain outcomes. You can see a discussion of this here: http://www.quora.com/Are-new-social-media-measures-and-metrics-like-Klout-and-PeerIndex-worth-integrating-into-existing-CRM-systems/answer/Azeem-Azhar

Thanks for continuing this debate, and look forward to speaking in person
Azeem
CEO PeerIndex

[Tom]

I wrote a similar piece entitled: Why Klout and Peerindex fail to measure your online reputation

http://invisibleinkdigital.com/social-media-2/klout-peerindex-fail-measure-online-reputation/

Azeem Azhar, the CEO of PeerIndex kindly responded to my piece. He paints a balanced picture that some people may have no need to know theirs or others level of influence. But is last words are quite pertinent in that as a service it can be used to save time looking up individuals.

[Brian Driggs]

One of the more commonly known "rules" of SEO says more inbound links to your site improves your ranking in the search results. So link farmers build sites full of links back to other sites they're trying to monetize. This is frowned upon because it's not genuine, organic effort. It's gaming the system.

Enter klout and the lot. One of the more commonly known "rules" of business is increased quality interaction with people improves business objectives. So people are ranked according to their "influence," making it easier for businesses to monetize their messages by greasing some influential palms. Is this genuine, organic influence? How is this not also gaming the system? How is this sort of thing different from link farms?

Influence, like brand, is a function of action. Influential people act. They do things. Scores like these serve to direct those unwilling to connect and engage with those who do. It's a model based on rating value, rather than providing it.

Any action taken to impact such a score is inherently insincere. Chasing influence is merely vanity.

[Gary Lee]

Thanks for continuing the discussion around Influencer identification and influencers in general. So far, the nascent industry for influencer tools has been fixated on "scores", and comparison of these somewhat generic scores to another's to determine influence. We struggle to see the world that way, since an influencer of one market segment may be extremely ineffective influencing another; and thus a "one size fits all" approach to influencer scoring does not hold up in the real world.

Properly done, Influencer scoring is never generic, never focuses only on Twitter and Facebook, and a single score cannot successfully be applied ubiquitously across the web like many tools are doing these days. In fact, done correctly, an individual’s influence score will vary by market segment – sometime wildly. For instance, an expert blogger with a high influence score on the market for network-connected television technologies at this week’s CES, may have virtually no influence if at all on the market segment for women’s beauty products.

To measure influence, you have to find the voices talking about the topics that matter to your target market, and you have to properly measure the influence those voices have by looking at their reach, authority and other heuristic-based algorithms.

That’s not a foreign concept for most marketing professionals, yet it is a foreign concept for many of the influencer tools today that hope to provide answers to professionals in marketing who really want to determine who the influencers really are for their market segments. Saying I am a 75, or 32, or 68, and comparing that to someone else who is a 64 really doesn’t tell me much as marketing professional trying to determine who to focus marketing efforts on. In fact, I think many of these tools are really about "the marketing of me", and a way to compare my score to others and find ways to boost my score accordingly.

Measuring influence requires tools that deal with topic-based research first to find “who is saying things my market cares about”, and then proper scoring algorithms to determine who amongst those voices is most important. To do it any other way is not recognizing how influence is really measured.

We're releasing such a tool on Monday of next week – our mPACT product. Would love to tell you more ahead of other reports you may write on this topic.

Gary Lee
CEO
mBLAST
www.mblast.com/mpact

[Matthew Thomson]

Great comments here but there is definitely one item I should straighten out. I commonly see others referring to the fact that Klout only has a monolithic score that doesn't take into account topical analysis. Basically, this is just...well...wrong.

We do topical scoring on every user. That's how we find the right people for Klout Perks campaigns like the Sacramento Kings (http://klout.com/blog/2010/12/the-sacramento-kings-have-klout/) or Disney's Tangled (http://itshardbeingperfect.com/2010/11/14/klout-perks-disney%E2%80%99s-tangled/). We certainly don't just take a random sampling of high Klout scores when choosing people for these campaigns.

The fact that we don't make topical data available to marketers speaks more to our interest in developing a relationship with our user base than simply turning them into a spam base for targeted marketing.

[Valeria Maltoni]

@Eric -- maybe that's due to your Facebook profile activity? From what I understand, the values compound between the two networks.

@June -- indeed, influence is not absolute and granted everywhere. And yes, it depends on what you're trying to achieve and how you go about it.

@Azeem -- thank you for taking the time to comment and for providing additional food for thought and links. The difficulty with the term you propose is the implication of superiority where there is difference... the overarching difficulty is, as you point out, the predictive nature of what brands and companies will expect. When you deal with humans, the most predictable outcome is that you will be surprised. So yes, you are attempting to put a picket fence around thought, and it is more akin to flow, or an ocean, with so many ingredients and species in it. Good analogy on your part. More than a debate, this is part of my lifelong work through and with may of the disciplines that study individual and group behavior, lots of science, yet not at the exclusion of the creative and contextual side of things. The danger with predicting outcomes is creating human billboards... and we know what happens to billboards.

@Tom -- thank you for stopping by and for sharing the link to your post and conversation.

@Brian -- "It's a model based on rating value, rather than providing it" well said. it doesn't take care of the value part of the equation, which is where the action connects.

@Gary -- more simply said, especially in the knowledge economy, no two individuals are interchangeable, no matter what the propaganda says. I'd take it a step further than “who is saying things my market cares about”. Much of what you say makes sense.

@Matthew -- and I know my Twitter persona confuses the heck out of everyone. In fact, people have a hard time pinning me down, period. I consider that a good thing. But that's a conversation for another day. Thank you for jumping in and clarifying your position. It does sound like people are interested in more transparency about the algorithm. Of course, they are already gaming the system anyway.

[Matt Hixson]

Hi Valeria-

Nice post as usual. I appreciate you continuing this conversation. Your post and all of the comments make good points. It seems like there are three main points - the continuing conversation on influence, some tools that are in the space and the backlash impact to Klout. The influence conversation will continue for a while because if it were easy this debate would not exist. It seems like the reality is that you can't apply todays understandings- web analytics or social reactions (RT of Like) or more complex things like sociology, social graph theory or other types of behavior studies. The answers are still emerging or evolving and are probably a mix of things we understand today and things we don't. It is a journey.

Because it is a journey if Klout does not lead that journey they will be a casualty or a non factor. When I look at Klout the business they either did something really smart or lucky. They created 1 simple score. People love to be ranked and that drove people to it. Most other sites have multiple things. Klout focuses you. I know there is more behind it but the main number is all most people look at and it is what has established their brand. I don't buy the number or the people I influence or am influenced by but I don't use it for anything more than a novelty.

I look forward to more conversation up to SXSW and getting a chance to meet you in person then.

Take care
Matt

[Adam B]

Where did you leave Soovox Social IQ?

[Jeff Katz]

Hi Valeria

Great post and comments. I head up product management at Twitalyzer and I would like to just make a quick comment or two about the notes about your use of our service

1) The profile that that we have (http://twitalyzer.com/profile.asp?u=ConversationAge) seems to match what is coming from Twitter (https://twitter.com/#!/ConversationAge)

2) The Network information is not just what you think is your direct network, but rather conservations where you included even if you are not directly part of that conversation (i.e. someone mentions you in a tweet) or are following that person

3) The reason why #kaizenblog tag was not picked up is based on the when you last used that tag in combination when your account was processed in Twitalyzer.

Feel free to shoot me an email at jeff@twitalyzer.com and I can through our analytics platform in more detail.

-- Jeff

[Paul L'Acosta]

Great points Valeria and amazing comments this conversation has spawned! For me, I don't believe in assigning a number to the people I'm interested in as it can easily pollute your mind and intentions. For example, a car dealer may pass on an extremely wealthy and responsible person just because of a score.

However, I can see a company trying to pitch a product and using someone's score to value who to use for their marketing efforts (kind of an interviewing tool for a spokesperson). Valid yet "cloudy" as @Eric said all the way up there.

For now, I stick to my believe that someone's influence is in the eye of the beholder (or on your Twitter stream).

Best of luck in 2011! ~Paul

[Rich Becker]

Valeria,

Thanks for the inclusion.

For anyone still confused about influence, I suggest they look back and think of their own lives and ground the topic in reality.

I was fortunate to have several teachers who shared lessons that stuck with me for life. They obviously had impact. Specific ideas, not all them, stuck.

Never once have I clicked on a link nor did I raise my hand or shout out when they first shared whatever it was stuck. Sometimes something they said became relative months or years later.

I don't even think Klout measures popularity or anything really meaningful. I do think they have created a machine similar to one once created by Sylvester McMonkey McBean.

[Aerin Guy]

Attempts at "Influence" can be bought, and one look at what the mommy blogger cohort has descended into is revealing. Once a community with it's own brand of clout, it's now largely a circle jerk of paid posts, badly disguised marketing reviews, and meaningless links to each other's sites. So yep, a highly Klouted mommy blogger has a lofty score, but is only influential within an increasingly circular club of exclusives, all fighting for PR dollars.

[Valeria Maltoni]

@Matt -- promise to connect at SxSW or even before. What you say makes a ton of sense, and of course it's where I'm headed myself. The person who simplifies wins. How it is with everything: processes, products, services, concepts, etc. So yes. However, to maintain the lead, it also needs to maintain the support while making sense. In other words, people may think it's not perfect, if they understand what it measures and how to contextualize it, they will get something out of using it as an indicator.

@Adam -- elaborate for the community? Teach us!

@Jeff -- thank you for taking the time to explain. I imagine you meant conversations ;) Maybe you used "conservations" as in stored/archived? For the hashtag, the reason why I was confused is it cached one I had used longer ago... I'll need to catch up with you on the tool itself. Thank you for offering.

@Paul -- it needs to be combination, who looks up to them, how they interact (tone, voice, content), as well as who follows them. I could have 20k bots following me, for all you know, unless you track and check. Indeed, influence is with the influenced.

@Rich -- certainly relevance (thus timing within situation) have a lot to do with it. We choose to be influenced when we're ready. Whatever Klout measures is not a reflection of what I could do... predictable outcomes.

@Aerin -- are we creating human billboards? What you describe has happened in other circles before online networks, of course. I worked very closely with sales groups my whole career and one thing they taught me is your credibility is everything in this business. It is.

[Geoff Livingston]

Great post, Valeria. And thank you for linking to me. This series of statements were particularly spot on: "there is no reliable qualitative information, as well as definite correlation to a desired change in behavior and actions in the community, even as in some cases, they give the impression that something may be happening or has happened -- a retweet, a link shared, etc.
It seems to still be related to those other impressions that mean so little to connecting trust and authority with relevance, and ultimately action."

And I look forward to your follow up post next week. Influence as a measure has a lot of kinks to be worked out. I also posted on influence and the social media bubble this week, and think that we can stand to look at some non-influence metrics a little more. You know, like business actions (not the watered down influence version).

[Dan Perez]

Valeria,
Terrific observation. For me, once I realized there were 20-something year-olds who didn't have a job but had Klout scores in the 70s, it made me ask myself, "Who the hell are they really influencing?". The reality is that the more time you spend tweeting, the better the chance your Klout score will go up. It's an even hollower number when you consider Lady Gaga has a considerably higher Klout score than Bill Gates. Greater influence? Don't think so.

People like numbers/stats. Those that fail to achieve any real accomplishments in the offline world, can still brag about a high Klout score online. It also makes for great blog fodder.

I do believe there is a place for services like Klout but only if you can leverage that score to your benefit. Folks like Chris Brogan and Brian Solis do this well. The rest of us? Not so much.

Better to build your influence where it really matters (and yields greater returns) with your family, your customers, and the man (or woman) in the mirror.

My full views can be read here: http://bit.ly/hub0eY

Looking forward to your follow-up post. Hug.

[Valeria Maltoni]

@Geoff -- as Lucretia said in her comment, there is a certain emulation effect in influence, from seeing people we aspire to be like - we all have our heroes and role models, some look at celebrity, what can I tell you? In that case, it is the influenced who determine what influences them, not influentials, if that makes any sense at all. More to come.

@Dan -- clever PR stunts have a way of fizzling out. Buzz comes and goes. Influence has enduring qualities for which those events have no match. They are playing finite games, and those will eventually come to an end. Just like in casinos, people will find that it's a trick, nobody bets and wins against the house. I digress. The people and agencies I'd hire as a client, and I have been there most of my career, are those who put their efforts into my business, not their fame. Which means the rest of us have plenty to do. Good thoughts on relationships and influence, thank you for the link.

[Taylor Davidson]

I love that you're driving the conversation on the topic of influence; it's an age-old topic with a new twist, powered by public communications. And with more and more about us showing up in the open public web, it's inevitable that we'll try to measure influence and rank people, it's a part of Western society. But so much about influence is intensely personal, that so much about it will never show up in any public form for a tool to measure.

Influence doesn't work as a currency for a lot of the reasons social capital doesn't work (although I'll say I haven't truly dug into the literature to delineate between the influence and social capital). For background thoughts about social capital, here's what I wrote a couple years ago: http://www.unstructuredventures.com/uv/2009/01/16/social-capital-is-not-new/

"Difficult to value.
Social capital is a poor asset: because social capital is held through a relationship between two parties, its value is subject to different valuations by each party and revaluation at any time by either party without the other’s agreement."

and

"Difficult to exchange.
Since its value is difficult to measure at any point in time and subject to unannounced, non-public and non-agreed revaluation, social capital suffers as a medium of exchange and can be difficult to cash in and exchange for other assets."

I'd posit that many of the same concerns about measuring and using social capital hold true for influence.

Looking forward to the next step in the conversation.