{{ searchResult.published_at | date:'d MMMM yyyy' }}

Loading ...
Loading ...

Enter a search term such as “mobile analytics” or browse our content using the filters above.

No_results

That’s not only a poor Scrabble score but we also couldn’t find any results matching “”.
Check your spelling or try broadening your search.

Logo_distressed

Sorry about this, there is a problem with our search at the moment.
Please try again later.

kloutFinding (or becoming) an influencer is often seen as one of the core goals for businesses utilising social media, and the search and measurement industry is rushing to fulfil this need.

Finding a great exponent for your brand who has a powerful presence on your social platform of choice and engaging them is a great way to get plenty of bang for your social media buck.

Increasingly leading measurement tools such as Klout are being seen as a good way to prove success in the social arena, with some companies starting to request a minimum Klout score as a deliverable when hiring an agency.

Unfortunately some of the systems that provide this measurement may still be relying on the wrong metrics, providing you with a skewed perspective on your audience and making them ripe for exploitation by unscrupulous users.

I read an excellent article over on RAAK recently, which detailed some of the ways Klout scores fail those using them, and was reminded of a post I wrote a while ago about best practices for Twitter.

I'd decided to try out a few different things myself and set up a dodgy ‘celebrity news’ stream as an example of how not to tweet. The stream amalgamates RSS feeds from various celebrity news sites and sponsored tweet sites and pumps them out at the rate of one every two minutes or so.

Obviously I also maintain a personal account where I tweet...well, any old rubbish really. But I do actually engage on a personal level there and usually garner a few retweets, particularly when I link to an Econsultancy article.

Given the vast difference in these two accounts it stands to reason that the ‘Real’ one should be far more influential than the ‘fake’ one.  On my account I reply, retweet, link out.

In short, I engage on a one-to-one basis and try to be a half decent digital citizen, while my bot account does none of these. You can shout at it to quit spamming you all you like; it doesn’t care, mainly because I only check it about once a fortnight.

Obvious, no?

Examine the numbers however, and things get interesting:

Both accounts now have a similar number of followers (There are obviously a lot of Justin Beiber fans out there). My own account has around 1200, the bot around 1150, putting us both in the top 1% of Tweeters, if not at the superstar level.

Of course, follower counts aren’t the best metric are they? Especially not if they consist largely of teenage pop fans and desperate Z-List celebs.

Unless you’re social measurement tool Klout.

While a recent change in Klout’s algorithm has given my personal account a slightly higher rating, if I run a quick search for the bot, Klout tells me:

You actively engage in the social web, constantly trying out new ways to interact and network. You're exploring the ecosystem and making it work for you. Your level of activity and engagement shows that you "get it", we predict you'll be moving up.

Which is of course, complete garbage. The bot doesn't engage with anyone, it doesn't network, it doesn't 'get' anything.

On the surface, this is a bit of fun, it lets you know if you’re doing things right or not and drops hints to help you improve. If you are using these figures for business though, it soon becomes apparent that the metrics being used are woefully inadequate, and possibly detrimental.

The problems are compounded when you consider that some businesses are now offering users special deals based on their social media influence.

The Las Vegas’ Palms Hotel recently went as far as setting up a ‘Klout Klub’, allowing influencers access to exclusive events, free drinks and deals at the casino.

On the surface this is a good idea, it means you’ll get influential people rating and reporting on your products. Or you’ll get a bunch of gamers with several bot accounts taking all the free drinks.

You can also tie Facebook data into your Klout account for added social power, but these figures make the crucial mistake of ignoring content.

Number of followers has very little to do with actual, tangible influence. I’m not going to buy a new lawnmower based on MC Hammer’s recommendation, and my Facebook account is an entirely different entity to Twitter

Frequency might be slightly more useful but again, it means that Klout is fairly easy to exploit because it doesn’t take the content of all those tweets into account. Sentiment scores are also fairly hopeless because “nice work breaking down again train company X” tends to be assigned a positive score.

In short, you can’t afford to remove the human element from your measurements. These measurements are fine for grabbing an initial feeling for a user, but you’d be just as well served by looking at their blog’s Alexa rank. These tools do have a function,but they are simply not mature enough to accurately deliver actionable data in this way.

If you are attempting to map your user base or reward influencers, you need to be prepared to spend time and effort on it.

Tools and apps simply cannot compare with human input. When finding influencers you want to engage, by all means look at the big numbers, but only as an initial benchmarking technique. Once you’ve found some likely candidates, you can spend time examining profiles manually.

To put this in context, ask yourself: would you hire a new head of marketing based purely on the amount of recommendations they received on LinkedIn? Of course not, you’d interview them, and if you want to succeed at social you'll need to be prepared for a lengthy vetting process.

Influencers can be useful, but only if they are contextually relevant, and it’s dangerous for businesses to rely on externally created numbers when attempting to leverage social influence.

Matt Owen

Published 6 December, 2010 by Matt Owen

Matt Owen was formerly Head of Social at Econsultancy. You can follow him on Twitter or hook up on LinkedIn.

203 more posts from this author

Comments (23)

Comment
No-profile-pic
Save or Cancel
Avatar-blank-50x50

Wessel van Rensburg

Thanks for the link Matt.

"you can’t afford to remove the human element from your measurements."

The same applies to an greater extent with social media monitoring. Sentiment analysis is still far from perfect.

While I think its possible to sharpen influencer algorithms - and that somebody - if not Klout will actually make a good stab at it, sentiment analysis is a notoriously difficult beast. Without some analysis by a person of the data, sentiment analysis is often meaningless.

about 6 years ago

Avatar-blank-50x50

Xavier Izaguirre

Awesome article.

It reminds me a brief analysis I had done. I had found you can gain a lot of Klout by chatting to people on a variety of subjects, while by doing so you won't be regarded as more influential on a specific subject.

You can then combine keyword density but then again you still need to factor in the quality of the tweets, or the quality of the mentions ( are people agreeing or disagreeing? )

In all fairness, algos can improve. Every piece of ciritiscism is a nudge for Klout to improve and so they will. Their problem is their marketing, as there is no way on earth you can count Klout as a standard on influence. Compound metrics are mostl;y useless and misleading by themselves.

Thanks for sharing

about 6 years ago

Avatar-blank-50x50

Ian Hendry

Matt, a really interesting article. It always worries me when things like Klout come along looking to work out what everything means with a smart algorithm. It is algorithms and the crappy results they produce that has people running from Gooogle and asking real people for advice and recommendations on social media instead as an alternative to sifting through thousands of pages of algorithm generated and often completely irrelevant answers. Trying to measure how "social" someone is through maths is overlaying Web 1.0 metric methods on the Web 2.0 Social Web. Anyway, that aside, I wanted to pick you up on your assertion that a stream of news on Twitter is "an example of how not to tweet" and "dodgy". Twitter is a news network rather than a social network. I find out about social media by following econsultancy and I do so for the news, not the social interaction. There are Twitter accounts for real people I have relationships with, but what's great about Twitter is that I don't need to have a relationship with someone to find out what they know or wish to share. I am aware that my second point here could be construed as contradicting the first (doing Web 1.0 stuff in a Web 2.0 world), but I see Twitter as delivering news over social channels in an ability that we can easily share and comment on that news, which is what makes use of Twitter in that way "social". Ian Hendry CEO, WeCanDo.BIZ http://www.wecando.biz

about 6 years ago

Avatar-blank-50x50

Will Rowan

aha!

I had just this conversation recently - and checked my Klout scores.

Taking Klout scores at face value is nonsense.

My personal account /thecustomer and the mass broadcast account for an online writing project /20todo which has been inactive for 2 years... the same Klout scores, give or take. 

& I suspect that businesses need to run both types of account - one to engage, one to inform.
The Inform channel's rate of news recycling will vary from industry to industry. Think of it as broadcast: impersonal marketing, distributed for folk to tune in to as & when they wish.

The engage channel is entirely different: highly personal, human-crafted. Pure one-to-one for customers who choose to engage.

Just because both channels run on Twitter doesn't mean they can't run alongside each other. Just don't confuse what each is there for!

Will

about 6 years ago

Matt Owen

Matt Owen, Head of Social at Econsultancy

Hi Ian, that's actually an interesting point, and I do broadly agree that for an organisation such as-for example - BBC News, a simple broadcast stream is acceptable, but I feel that aggregator feeds that only tweet RSS feed content simply add unneeded noise and utilise other people's content to broaden their influence. We've found more than a few twitter accounts that broadcast the Econsultancy feed verbatim and do not engage with their audience in any other way, which I believe shows a lack of understanding (and it's worth considering how you'd approach the practice if it were someone republishing all your blog content every day...). It's fine to broadcast news, but not to ignore any other engagement. Twitter accounts should be willing and able to respond directly to their followers and carry out conversation -Twitter is indeed a news network, but it is also extremely social and needs to be approached as such.

For the record, I'm not inherently opposed to anything that Klout do, as I believe there is a use for tools that provide broad figures as a means for benchmarking, but I do feel it's a mistake to put too much emphasis on these measurements. It's an extension of looking at follower counts as a means of determining influence, and doesn't account for more useful metrics or influence on other platforms or offline. I have no doubt that algorythms will improve, and we're already seeing some interesting developments (particularly concerning semantic technology), but marketers need to realise that these figures are fairly easy to game and don't provide an accurate picture.

about 6 years ago

Avatar-blank-50x50

Azeem

Hey Matt

So I think two different things are going on.

And they are best viewed as the difference between running a google search and having an analyst present you some data.

Any algorithmic scoring system is (for a while at least) going to be the equivalent of a Google search -- narrow a search space of billions of people to a few dozen you can process. 

Woe betide the company who runs their business ont he first result google brings up.

Likewise for any platform which identifies people by various metrics.

One way of narrowing the search space is also to see how resonant people are in particular topics -- it's always been an axiom of PeerIndex that an authority in pharmaceuticals may not be relevant in football or vice-versa. 

And the result is that you identify people who matter a lot in a niche of a community, even if their 'global score' is less impressive.

about 6 years ago

Avatar-blank-50x50

Simon D

Interesting article. These are the types of social media case studies I'd like to see more of: ones that challenge lazily accepted conventions! 

However, I think you're wrong to confuse influence with engagement. A century's worth of advertising and PR using traditional channels has proven engaging on a personal level is not a prerequisite to influencing consumer decisions. You haven't provided any evidence that your personal account is any more influential than the fake account. You just seem to assume it is because you're more interactive on it. 

I understand why there is this confusion. It's incredibly difficult to gauge influence in a meaningful way - and doing it manually doesn't really help much. Sure, you can judge the relevance of the voice's audience, subjectively decide whether or not their content is right for your needs, and see how actively they participate with other social media users, but does all this really add up to influence? 

Trying to find previous examples of a voice's influence 'in action' is tricky and inconsistent. Retweets and click-throughs seem a good start, but are not necessarily dependent on engagement and therefore don't fit with your argument. 

I think the idea of an 'influence' metric needs to be dropped. It's inaccurate and vague. Instead, a subset of different measurements should be used for their own purposes; a mix of qual and quant. Reach is what most companies are looking for. Engagement is what you're interested in. Audience Relevance gives a different perspective (though for some purposes may be less important). Content could be another piece of the puzzle - useful if you're looking for particularly creative or entertaining people to participate in a campaign. These are just four possibilities. Does this all together add up to influence? Possibly. But also possibly not. We just don't know, so let's not pretend we do and instead start using terms that make sense. 

about 6 years ago

Avatar-blank-50x50

Matt Kelly

I agree that Klout scores and the algorithm behind them are clear as mud but....

What's a better solution to answer the proverbial question your boss is going to ask you about who your latest social media initiative is performing?

As a guide to determine reach, influence, and most importantly who will take action to potentially increase your visibility Klout is a fairly good tool to help get some direction in the swirling waters of social media.

Matt Kelly
Business Development
ideffect.com 

about 6 years ago

Avatar-blank-50x50

Tim Wilson

Great points, and I love the way you set up the experiment! Any chance you've thought about setting both accounts up in Twitalyzer? One of the fundamental ideas behind Twitalyzer is that it provides multiple different calculated metrics, with the idea that different metrics are more appropriate for different types of accounts (a "personal brand" account vs. an "agency" account vs. a "retail brand" account...as well as what the account is trying to accomplish). But, those metrics are also intended to be able to counter-balance each other -- if you game one, another will suffer. I'm a Twitalyzer user -- not in any way directly affiliated with the company.

about 6 years ago

Avatar-blank-50x50

Megan Berry

Hey Matt,

I wanted to stop by and thank you for the post and feedback. I'm the Marketing Manager here at Klout and we are working very hard over here to constantly improve our algorithm. 

First of all, I want to say that follower count is actually a very small part of our algorithm. I even have a blog post on the Klout blog that points out the same flaws with follower count you mention (http://klout.com/blog/2010/06/why-follower-count-doesnt-measure-twitter-performance/). We are focused much more on engagement. Sometimes bots do end up getting @messages and RTs even when we believe they don't deserve them. This is something we are working on solving.

Secondly, I agree that relying solely on one number -- no matter the number -- is likely not a good idea for a business trying to engage with someone. However, our belief is the Klout score, coupled with either other insights from the company or from the Klout profile, can take a business (or person) a long way towards understanding the person in question. 

Thanks and feel free to reach out with more feedback.

-Megan Berry

Marketing Manager, Klout

@meganberry

megan@klout.com

about 6 years ago

Matt Owen

Matt Owen, Head of Social at Econsultancy

Hi Megan, thanks very much for your comment. I should say first that while I'm agreed that there are a few tweaks to be made, I'm really not opposed to Klout as a tool and it can actually provide some very useful information, all metrics do have their uses. My real point here is that there does seem to be a lack of understanding by many companies regarding the actual application of these figures, which does leave the system open to abuse, particularly - as in the Palms' case- when there is a direct reward for the 'high scorers'. Anyone can build a large following if they go about it the right way, and anyone (well..nearly anyone) can become an 'influencer', but who they are influencing and how needs to be a deciding factor. It really is a case of strategy and granular research rather than simply concentrating on big numbers. The trouble occurs when companies start to ask (as indeed they are now) for increased scores as a deliverable. This is exactly the same as asking for a larger number of followers. Given a few weeks it's fairly easy to build a large following, but whether or not any of your followers are of use to you (or indeed, if you are of use to them) is quite a different matter. Thanks again for getting in touch. Matt.

about 6 years ago

Avatar-blank-50x50

Michelle Wohl

Hi Matt - Great article. We provide a software service that helps hotels manage their online reputation. In the hospitality industry or any industry that really values customer service, it is hard to make a case for treating people differently based on klout. If I am a GM and a customer tweets about a problem he's having, I want to make sure that I acknowledge and solve the problem... whether he has klout or not. My focus needs to be on the customer and winning his loyalty, regardless of how many followers he has.

about 6 years ago

Charlie Osmond

Charlie Osmond, Chief Tease at Triptease

Hi Matt,

Great bit of investigation. 2010 has been the year of people working out that you can't use a social media monitoring tool without a good amount of human anlytics.

You might be interested in a report we just published on influencers. we compared 9 of the top social media monitoring tools against one another to see which was best at finding influencers.We didn't bother with Klout for obvious reasons.

We focussed (as Azeem suggests above) on a particular cateogry - influential Mummy bloggers on organic baby food. WIth some interesting results.

The report is free: Finding Influencers using social media monitoring tools

Charlie

about 6 years ago

Avatar-blank-50x50

Steven Hughes

It's all perception. If organizations and people think Klout counts, then it counts. If it didn't count you wouldn't be writing an article on them as no one would care. Furthermore, a Las Vegas casino wouldn't be giving the house away if it didn't account. Your points about how Klout scores are definitely valid, but that is separate from if it counts. Article should have been named, "Klout scoring has some definite holes, but it does count."

about 6 years ago

Neil Warren

Neil Warren, Publisher at 2N Media Ltd - ModernSelling.com

It’s probably just the same specs I’ve been wearing for the last 25 years (my Selling Specs), but I can’t help thinking that this is, STILL, the Great Divide between sales and marketing, people versus machines/collateral, B2C v B2B, and so on.

The same questions were asked along the lines of…

Is it better to pay a higher cost per thousand in a niche/trade magazine, or save up and zap them all in a national newspaper?

When I wish to mail “Senior Sales Decision Makers” (my example), should I pay more to discover which ones actually manage a sales team, or just take 150,000 cheap ‘n’ cheerful versions to cover all possibilities?

Is it a good idea to outsource and offshore my telemarketing effort, so that the labour and call costs are cheap, and then I can phone everyone in the UK with a script (or even a bot recording) to see if they want to spend £10,000 - £20,000 on a new kitchen?

Email lists, they cost very little more to actually send, so the maximum number for the cheapest price is a good thing, right?

Ask any sales person what value they attach to the last “lead” they got from marketing, and you’ll see what the true value of those are – whether measured by Klout or anything else (and I’m not familiar with your service Megan, so am not casting aspersions any more than I’m endorsing).

I’m with Ian (Hendry) in that I can easily see how the expectations of both the Tweeter and the Follower(s) can indeed be limited to news distribution. I would have no hope in using Twitter at all for actual continuing dialogue, not least because I’m anyway “lamentably verbose”, but also because my more detailed B2B CONVERSATIONS with clients and prospects very quickly exceed a bullet point. So the meaty content has to be elsewhere than that platform (e.g. like here, which I’ll Tweet when I’ve posted). If you can exchange useful information in 140 characters though, then of course do – it really depends on why people (and it’s human beings we’re talking about here) are following you in the first place.

Other examples that have struck me over the years were the Dell following – I assume mostly to pick up on the latest discount (and then un-follow or ignore until another PC/laptop is needed) or that chap in the States who advised his followers when the taco van was coming to their street – but I guess most didn’t want to “talk taco” day-in day-out.

The point is that if you think of an “audience” in terms of each individual prospect or customer (as most sales people are obliged to do) you are less prone to trying to “abuse” them in any way, than if you think like a “marketing machine” and really do still believe that getting the message out (heard or unheard, seen or unseen, relevant or irrelevant) is the goal.

about 6 years ago

Avatar-blank-50x50

Rene

I've had a post about why I view Klout as largely irrelevant in draft for a couple of months (sorry Megan!) and now Matt you've gone and conducted a study, generated some  actual evidence and blown what I had planned out of the water! A salutory tale in writing and posting when you have the idea.

Great post by the way, in agreement with much of the sentiment!

about 6 years ago

Avatar-blank-50x50

Gary Lee

Interesting post on a very imporatnt topic -- how to find influencers and properly identify them.

Your last comment "Influencers can be useful, but only if they are contextually relevant" is the one I believe most important and worthy of study.   Marketing professionals for years have been dealing in "contextually relevant" mapping of influencers when determine who to focus their PR and marketing efforts on.   Good marketing folks look first at context -- who is writing on what topics and things of relevance to their target market segments, and then based on this, look at who is the most influential.    This is not new.  It's been done for years, it's just a lot harder now to do due to the volume of conversations going on daily online.    For more thoughts from me on this, see:   http://blog.mblast.com/mbwordpress/?p=166 and also a video we produced on finding influencers -- including the need for contextual / topical relevance:  http://www.mblast.com/mk/solutions/marketers/mpact/mpact_landing.aspx

about 6 years ago

Neil Warren

Neil Warren, Publisher at 2N Media Ltd - ModernSelling.com

Yes that is a very interesting historical perspective Gary, on your blog.

Coming up to the present day though and it still looks as though we're trying to keep PR, Advertising, Marketing and so on in the loop - as well they may need to be for "consumer" and mass-market influence. But what is anyone suggesting we "badge" or think about a directly employed sales representative, who is him/herself charged with becoming a "trusted advisor", "go-to-guy/gal" or, even better, a "sales architect" - like those IBM sales people who blog? And is the "loop" then not needed - if buyer can now best get directly in touch with seller, without needing "the media" (and nurturing marketing outputs) as much?

Should "Marketing" in fact then be trying to feed the relevant "collateral" into the sales conversations where, typically or ideally, each of 10 to 10,000 "Sales" staff will anyway be the key influencers for each of their 50 to 5,000 customers and prospects (followers, tribes etc..)?

Procter & Gamble training Customer Service Representatives to be out and about engaging with the (focused) masses was the nearest I have seen to the B2C version of this - by the by.

about 6 years ago

Avatar-blank-50x50

Gary Lee

Neil: Your points are well-taken, and I think the methods we're advocating would apply to both sales and customer service professionals as well. It's important to identify who these key influencers are based on their topical relevancy, and then determine within the organization is best to insert themselves and add value to the conversation. What I believe you're describing in an organization decision -- P&G has one mindset. Other companies view the "ownership" of the customer conversation differently. But the tools needed for properly identifying influencers apply to anyone by and large.

about 6 years ago

Avatar-blank-50x50

Chris Gillett

Really insightful and detailed research Matt.

Klout has been the source of some debate here @3seven9, and it's good to have some facts available to refer to. I know Klout themselves have said they are working on updates so it will be interesting if they can correct their monitoring, which largely seems to rely on volume only (although they have been quick to point out even Bots get retweets and replies which count as 'engagement')!

What i find most bizarre/annoying is Hootsuite's reliance on and promotion of Klout itself. The only way to filter Searches are via Keywords or Klout! There are much much better monitoring and influencer tools out there and having the option of using/integrating some of them is essential now Hootsuite moved towards a paid service.

As a footnote: a lot of those influencer tools are reviewed by FreshNetworks as Charlie has already stated, and i can highly recommend it - very useful; http://ow.ly/3lnvA. Thanks again for your hard work on the research here!

Chris @3seven9

about 6 years ago

Avatar-blank-50x50

marK clayson

Loved the article and the discussion it started!

I tend to take my Klout stats with a pinch of salt unless they show me in a better (good) light! Human nature dictates that we only want the good in us to be seen.

When all is said and done, a real user with positive human reactions will always rise to the top. Just like being at a party - the fakes annoy us all.

about 6 years ago

Avatar-blank-50x50

Derek Skaletsky

Great article and you are bringing up a point that is finally getting a lot of heat and discussion -- the notion that influence is contextual (Matt over at Recessive Networks wrote a good post on it recently: http://ressivenetworks.com/2010/11/influence-in-context/).  I'm with TRAACKR and we've always believed, first and foremost, that influence is contextual.  Measuring someone's Relevance to a particular topic is highly correlated with his/her influence in that discussion.  Relevance is actually the most important metric we use in our scoring of a person's influence.    

With that said, it is always interesting to find that there are people with strong, loyal audiences who sit on the periphery of a topic/conversation who could also be influential in that conversation if they decide to focus on it more.  So, when we're talking in the context of influencer identification/discovery (which we do every day), it's important to understand the subtle distinctions.  You can check out our blog for more discussion on the topic.  

Also, check out Matt's recent post where he talks about how Klout has said they don't measure influence:  (http://ressivenetworks.com/2010/12/klout-admits-it-doesnt-measure-influence/) 

Keep up the good discussion!

almost 6 years ago

Avatar-blank-50x50

JD

Alexa rank? Really? I've had four self-hosted blogs in the past nine years and not once has Alexa even come close to accurately ranking my site's visitors or demographics.

As for Klout, they show me to influence people I haven't engaged with for months. They show me to be on a hundred lists when I'm on 400, and they don't count RT's, only the official "Retweets" of Twitter. Three days ago I was a "Thought Leader" and today I'm a "Specialist" -- a seeming demotion in rank although yesterday I was retweeted close to 70 times.

On the positive side, they are responsive on Twitter and will review your Klout account if asked.

Still, I think people take these ranking tools too seriously. The best feedback comes from other users, not machines.

almost 6 years ago

Comment
No-profile-pic
Save or Cancel
Daily_pulse_signup_wide

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Daily Pulse newsletter. Each weekday, you ll receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.