Measuring Online Influence

The Official Klout Blog

Archive for the ‘measuring influence’ Category

A Year of Klout

Wednesday, December 28th, 2011

Moving fast, making friends, breaking things and trying to change the world. I would say that 2011 had all the ingredients of an amazing year for the Klout team.

With all that has happened, it feels impossible that January was only one year ago. I remember being excited but nervous heading into 2011. We had just closed a round of financing with Kleiner Perkins and Greycroft. The pressure was on to evolve Klout from interesting novelty to meaningful business. We were also at a crossroads in terms of how we wanted to think of our business. The obvious path was to be a data-centric company focused on helping businesses target influencers. This didn’t feel right though and we made the call to continue our consumer focus. We felt that by empowering people to understand and leverage their influence we could have the biggest impact. We wanted to unlock the world’s influence.

Our goals coming into 2011 were to be “the standard” and to make the Klout score matter. To be the standard, we wanted to be recognized for having the highest quality data and to have the Klout score distributed everywhere. From using Klout to quantify a politician’s traction, to credit cards integrating Klout to reward customers with loyalty points, the Klout score became the industry standard. In fact, distribution of the Klout score grew from about 1000 developers in January 2011 to over 5000 partners now. Even more exciting is the amount of API calls, which grew from about 100 million per month last January to over 7.5 billion per month now!

The bigger challenge for Klout in 2011 was to make the Klout score meaningful and relevant to everyone. In May, we relaunched Klout.com and saw a huge surge in engagement and user registrations. Soon after, we released +K which allows users to explicitly recognize their friends’ influence. +K has been a huge success with over 50,000 +K’s generated daily. We reintroduced our much improved topic pages and also added LinkedIn, Foursquare and Google+ to the algorithm while beginning to analyze YouTube, Tumblr, Wordpress, Instagram, Flickr and other networks to provide a complete picture of an individual’s influence.

Klout Perks also made a huge impact in 2011. Over 250,000 people were rewarded for their influence with amazing Perks from brands like Spotify, Red Bull, Hewlett-Packard, Turner, American Express, Chevy and Microsoft. I am particularly proud of the Klout Perks program. To have top brands recognize the impact not only of social media but also the power and voice of the individual was exciting to see. These brands followed our Klout Perks Code of Ethics and simply let influencers experience their products without knowing what the influencers would say and getting no private information about the participants. This past year we also redesigned our Klout Perks packaging, so be on the look out for the orange Klout Perks box because we are looking to take Klout Perks to the next level in 2012!

With all of our successes, 2011 wasn’t without some major challenges and screw-ups. Measuring influence is a monumental challenge and as the social media landscape continues to evolve, our algorithms need to evolve with it. In October, we launched the biggest algorithm change in the nearly four-year history of the company. In planning for this change we thought we had our bases covered in terms of transparency and communication. But it’s clear we didn’t do enough. The reaction was a great kick in the pants for the team and we appreciate everyone who cared enough about their Klout scores to #occupy us.

Privacy is another area where we learned some tough lessons in 2011. The nature of social networks and the data Klout analyzes makes privacy a consideration in every decision we make. In an effort to be more transparent about the data Klout analyzes, we launched the “Understanding Klout” portal. Our goal for 2012 is to be seen as a leader with respect to privacy.

Through the ups and downs of building a startup, my favorite part was getting to work with so many amazing people. From our investors to customers and partners we were lucky to have so many brilliant minds helping us. The Klout team grew from 15 people to more than 50 and I have never met a more humble and passionate group. I also want to thank our users for inspiring us to push harder and do better. We appreciate your support and patience and we look forward to helping you unlock your influence in 2012!

Posted in measuring influence | View Comments

The Most Influential Topics of 2011

Thursday, December 22nd, 2011

In 2011, the world’s major events unfolded on social media. This is true across the board, from the Occupy Movement to the Netflix scandal to product launches. We took a look at the most influential companies, locations, people, music genres and tech products of 2011.

Instead of just looking at the volume of mentions, we focused on how much response and engagement these subjects get. When influencers shared content about these topics, the world listened and responded.

Companies: For some companies, making the top list reflects their beloved product and people (see: Apple and Steve Jobs), for others, it reflects scandal (see: Netflix). We’d also like to call out SoundCloud for making the list; we suspect their star investor, Ashton, might be one of the reasons people were talking about them.

Locations: Contrary to the idea that “What happens in Vegas, stays in Vegas,” influencers loved sharing about Las Vegas and it took the top spot for location of the year. This likely has has to do with the amount of top conferences and events held there every year and, let’s be honest, probably with some serious partying there as well.

People: The top people list reveals how much music is an essential and engaging part of the social media landscape, the top four spots are music artists, even beating out Barack Obama for the most engaging topics of the year.

Music Genres: Turntable.fm, SoundCloud, and Spotify all contributed to Dubstep gaining the number two spot in music genres, coming out ahead of everything but Hip Hop.

Tech Products: When we turn to tech gadgets, it’ll come as no surprise that Apple owns this list, taking over the top four spots. The iPhone won the coveted top smartphone spot, with Android (we grouped the individual models together), Palm Pre, and then Blackberry (they would have been number 12 so they’re not on the graphic below) following behind.

What were your favorite companies, locations, people, music genres, tech products of 2011?

Posted in measuring influence | View Comments

December KloutChat on Topics!

Wednesday, December 7th, 2011

#KloutChat is a regular event on the first Wednesday of every month.

This month’s KloutChat is this afternoon and we’ll be talking about our recent Topics improvements and what else you’d like to see us add.

#KloutChat

Wednesday, December 7th at 3PM PT / 6PM EST

Join us for our first chat on Stanzr at stanzr.com/kloutchat

Today’s chat will take place on Stanzr! It’s a great new way to organize chats and also allows you to join whether you’re on Twitter, Facebook or LinkedIn. All you have to do is go to http://stanzr.com/kloutchat at 3PM PT, sign in on the upper left and join the conversation. You can also check it out now to see the questions we have planned for the chat.

We’re looking forward to a great convo! Remember, if your question is specific to your account it’s best to go to support.klout.com and get it answered there. Thanks!

Posted in measuring influence, social media | View Comments

Do You Have Google+ Klout?

Tuesday, November 22nd, 2011

Our mission here at Klout is to help you understand and leverage your influence. As we add more networks we can more accurately measure your influence. Today, we are excited to announce that Google+ is now a part of the Klout Score!

Since users started connecting their Google+ to Klout in late September, we have been hard at work building a model for G+ influence. With its explosive growth and innovative features, it’s a notable addition to the Klout score.  From the get-go, Google+ came with circles, which set boundaries for shared content.  As on Facebook, you can Like content from other users (called a “+1”) and comment on it. People validate your content by +1’ing it.  They can also Reshare it — which will spread it further through your own circles as well as spreading it to their own. Since it launched, Google+ has seen early adopters crossing over from Twitter and Facebook, especially in the tech industry.  The Reshare, which is in some ways similar to a Retweet, allows Google+ users to comment and discuss topics of interest similarly to Facebook, while having an immediacy of Twitter.

We are currently measuring influence for users who have already connected their Google+ accounts to Klout. Of those users, 62% are active on Google+ and therefore should see it affect their Score based on their ability to drive action on that platform.

Google+ and your Score

If you are active on Google+ and have connected your Google+ account to Klout, you will see a Score increase. Regardless of activity level, no user with Google+ connected will see a Score drop. Similarly, we do not penalize users who do not connect Google+ accounts. We measure influence on Google+ by analyzing public posts.

Distribution of Google+ Influence

The graph below shows the distribution for active users before and after adding Google+ to their Scores. In the graph below, the x-axis shows ranges of Scores. You’ll note that users active on Google+ tend to have scores between 30 and 60, which is much higher than our overall average Score of 20.

The median Score change for active connected users is 0.4742 and the mean Score change is 2.1577. We did see some users who have really embraced Google+ as a platform saw Score jumps of up 40+ points.

Comparing Google+ to Twitter and Facebook

Google+ has a hybrid of characteristics from both Twitter and Facebook. Facebook tends to be a tightly connected network, where users are connected to each other based on mutual agreement. You can connect to someone on Facebook only if they have accepted your friend request or vice versa. Twitter is a more loosely connected network because it is possible for you to be connected to someone whom you may not know. For example, you may follow Barack Obama, though he does not know you personally or follow your tweets. Google+ incorporates elements of both models with the use of circles, and is therefore more tightly connected than Twitter, but not as tightly connected as Facebook.

We compared the behavior of users on Google+ based on their other connected networks. The first plot shows how the distribution changed for users who are active on Twitter and Google+ only, the second shows the same for users who are active on Facebook and Google+ only, and the third shows the same for users who are active on all three networks. The last (with all three connected) is the largest group. One observation from these plots is that most active Google+ users are also active on Twitter and Facebook, which is why most users fall in the 30 – 60 Score range, and why you see a smaller change in the last distribution.

Comparing Social Actions on Google+, Twitter, and Facebook

Key similarities:

  • A comment, like or +1 on either Facebook or Google+ is a sign of engagement between the user who posted the content and the user who commented / liked / +1ed it.
  • A +1 on Google+ is similar to a like on Facebook, in the sense that both are validation mechanisms for “approval” of content.
  • A Reshare on Facebook or Google+ or a Retweet on Twitter both propagate the original content to a wider audience.

Salient differences:

  • A comment or a Like on Facebook may cause your content to appear in the news feed of mutual friends. On Google+ a comment or +1 surfaces your content not just to mutual friends but to anyone who has added you to their circles.
  • +1s on Google+ may also cause your content to appear in Google search results.
  • Due to the larger size of the network, a Retweet on Twitter may propagate your content to a wider audience than a Reshare on Google+ or Facebook. On the other hand, a Reshare on Google+ or Facebook may help you engage with an active audience, even if the size of the audience is smaller.

As part of the process of building this scoring model, we found comparisons with other networks helpful, but ultimately we look at each network holistically to create a scoring model.

Google+ Stars

We took a closer look at the “movers and shakers” on Google+ whose Scores increased significantly in this release. There are two interesting categories whose Score went up:

  • New Voices: Users who “jumped” on Google+ and found their voice there.  Their Score was low to begin with, and we see some impressive rises here.
  • Growing Influencers: Users who were already influential on other networks, who increased their engagement through Google+

New Voices

Erica Joy (G+ account) works at Google TV and her Klout Score had risen by a whopping 62 points! She shares content predominately about politics and food, getting lots of engagement from her audience.

Tamara Pruessner (G+ account) is a nature photographer and uses Google+ to showcase her photos. She’s found an engaged audience who follows her work and that’s increased her Score from 11 to 67.

Growing Influencers

Thomas Hawk (G+ account) is a photographer, regularly appearing on http://www.thisweekinphoto.com/.  He uses Google+’s excellent photo-sharing features which allow viewers to see all photos from posts conveniently. He is also an advocate of Google+ and connects photographers to the community.  An already high Klout score of 69 rose to an amazing 82.

Tom Anderson (G+ account) is the founder of MySpace.  He is “happily retired” and a power social media user. He took to Google+ with gusto, as many early technology adopters. He discusses topics such as gadgets and modern art, generating many comments from his numerous followers — he’s been added to circles more than 400,000 times!  Tom’s Klout Score has increased from 64 to 77.

Jason Calacanis (G+ account) is a serial entrepreneur, angel investor, and writer.  He shares a lot of photos about political issues, as well as posting about his startup activity, new apps, and related events.  Like Thomas Hawk, he runs a This Week show — but for startups instead of photography. His Score jumped by 11 points.

The same trends are reflected among folks at Klout. Out of about 50 people, only 4 saw a bump. For instance, Megan Berry (G+ account) rose by 3 points, while Derek Wollenstein (G+ account) rose by almost 4 points, proving that being a tech guy pays on Google+, even if you only talk politics there!

In summary:

  • Active Google+ users on Klout will see an average Score increase of 2.177.
  • Active Google+ users tend to also be active on Facebook and Twitter.
  • Google+ allows for close interactions like Facebook, while retaining the ability to share widely like Twitter.
  • Users who see a big jump in score from Google+ generate a lot of discussion and engagement with focused, high-quality content.

We’re excited to add Google+ as the latest of our networks and would love to hear your feedback. Also, check out our own Google+ page!

Update 11/30: We recently resolved an issue with G+ data collection and have updated accounts that were affected. Today, some users will see an increase and some users will see a decrease as we take into account the new data and the distribution changes. Active Google+ users on Klout will see an average Score increase of 2.1636.

Posted in announcements, measuring influence | View Comments

We Value Your Privacy

Sunday, November 13th, 2011

Privacy on the social web is an incredibly complex issue that we approach with the utmost seriousness. Our mission is to help people understand their influence and to get the most out of it. With that goal in mind, it is critical that we are model citizens in this space and do everything we can to respect the privacy of our users. That said, like Facebook, Google, and nearly every other company in this space, we are working hard to figure this out, but will not always get everything right. Recently, we erred in creating profiles for registered users’ Facebook friends who had recently interacted with them and did not have private accounts. The public interactions were used to show which Facebook friends that registered user was influencing. While this is ok per Facebook’s policy, the data returned to us about friends does not include age information and when we realized that accounts for minors had been created we rolled the changes back.

We will always be vigilant in working with the platforms (Twitter/Facebook/etc), our legal counsel, and the community to do what’s right here. We messed up on this one and are deeply sorry.

I think it’s important to be specific, so here is how Klout thinks about and has always thought about social data:

  • Klout analyzes public data to measure a person’s influence. The best way to think of this is in relation to how Google analyzes public websites to generate PageRank.
  • Klout respects the privacy settings of all the networks it measures. If you have a private account on a social network and you have not explicitly given Klout access to your data it will not be analyzed.
  • If you are creating public data but do not want it measured by Klout you can opt-out by going to the privacy page.
  • Klout has no interest in understanding the influence of minors. We are working with Facebook and Twitter on this, as well as building our own safeguards to make sure this does not happen.

We look forward to continuing to work with the community and our partners to protect your privacy.

Posted in measuring influence | View Comments

The Life of a Tweet

Friday, November 4th, 2011

How long does your content last? We took a look at the life of a tweet for influencers with different Klout Scores. We found influencers with a Klout Score above 75 have a half-life up to 70 times longer than those with a Score between 30 and 70. Messages from these high-scoring individuals stay active and meaningful for a longer time, illustrating their influence.

Tweets created by users with a Score under 30 have a longer half-life but a much lower overall volume of retweets. Those with a Score between 30 and 70 get their messages spread out to the network within the matter of minutes, but are not as adept at having their messages last longer within their network as the highest scoring Klout users. Of course, unsurprisingly, we also noticed a growth curve where online influencers with higher Klout Scores get their messages retweeted by more users.

Check out our results below or see the full image. On top is the half-life of users with different Scores. Below, we see an increase of retweets for users with higher Klout Scores.

Behind the scenes:
We used about a week’s worth of retweet data to include users tweeting on weekdays and also those who use twitter occasionally mostly during the weekends. From this data set we filtered out those retweets which originated earlier than our sample timeframe. We also cut retweets where the original tweet was created in the last 24 hours of our sample data since a significant number of their retweets could fall out of our data set. Dealing with hundreds of millions of retweet message records we used map/reduce to group these messages by the original author and calculate aggregated information about all the their re-tweeted messages.

Let us know what you think!

Posted in measuring influence, social media | View Comments

November KloutChat!

Tuesday, November 1st, 2011

I’m pleased to announce this month’s KloutChat is tomorrow! We’ll chat about our new scoring release, some new privacy features, and share exciting upcoming features!

#KloutChat

Wednesday, November 2nd at 3PM PT / 6PM EST

We’re looking forward to a great convo! Remember, if your question if specific to your account it’s best to go to support.klout.com and get it answered there. Thanks!

Do you have any questions you want to make sure we address? Post them in the comments and we’ll try to add them to the agenda. #KloutChat is a regular event on the first Wednesday of every month.

If you’re unfamiliar with Twitter chats, check out the article “What is a Twitter Tweet Chat?

Posted in measuring influence | View Comments

A More Accurate, Transparent Klout Score

Wednesday, October 26th, 2011

Today we’re releasing a new scoring model with insights to help you understand changes in your influence. This project represents the biggest step forward in accuracy, transparency and our technology in Klout’s history. Joe shared the full vision behind these changes in his post last week.

Influence is the ability to drive action and is based on quality, not quantity. When someone engages with your content, we assess that action in the context of the person’s own activity. These principles form the basis of our PeopleRank algorithm which determines your Score based on:

  • how many people you influence,
  • how much you influence them and
  • how influential they are.

We analyze 2.7 billion pieces of content and connections daily. Reaching this scale, we’ve introduced significant upgrades to our platform, allowing us to handle this explosive growth. Now, we can add more networks and other sources of your influence much, much faster.

Insights help you understand why your Score changed. Each day, you can see which subscore and people in your network caused that change. You can also view insights on your friends’ profiles.

These changes are a significant milestone in the Klout Score’s evolution and you can continue to expect more improvements in the future. As always, your opinion is very important to us and we’d love to hear your feedback.

How will this affect my Score?

A majority of users will see their Scores stay the same or go up but some users will see a drop. In fact, some of our Scores here at the Klout HQ will drop — our goal is accuracy above all else. We believe our users will be pleased with the improvements we’ve made. Below is a distribution of the Score changes. You’ll note large decreases in Score are rare.

Posted in announcements, engineering, measuring influence | View Comments

A New Era for Klout Scores

Wednesday, October 19th, 2011

More than three years ago the Klout Score was born in my bedroom in New York City as a way to make sense of the noise I was seeing in social media. I could share my opinion about anything, instantly, with the people who trust me and the data was available to measure my impact. Fast forward to today and we now have over 3,500 companies using the Klout Score to reward influencers with Klout Perks, give better customer service, reward loyalty, recruit, and much more.

The biggest change in the past three years is that (thankfully) we have people way smarter than me spending each and every day improving the algorithms that calculate the Klout Score. I am incredibly proud of the work the team has done and I am excited to announce the biggest improvement to the Klout Score in our history is launching next week.

People Rank
We’ve often thought of what we’re doing as a form of PeopleRank and this is a giant step in that direction. We’ve improved the stability and accuracy of our scores. Furthermore, our subscores have always been an important part of Klout. This update will make them more clear and make changes easier to understand.

True Reach
True Reach is the number of people you influence. It is a real number of people we find by looking at the impact you have on your connections. We analyze over two and a half billion connections and pieces of content every day in order to accurately gauge who is in your true reach.

Amplification
Amplification is how much you influence these people. We analyze how many people in your potential audience act upon your content. We take this a step further and understand what an influence signal means in the context of that person. For instance, if I rarely like or comment on anyone’s posts, but choose to do so to yours, that is more meaningful than if I like 60 posts a day. Amplification indicates the effect you have on your audience.

Network Impact
Network Impact is the influence of your audience. This is on a 1 to 100 scale and indicates the influence level of people who engage with your content. It’s not just about how many people you reach, it’s about getting your message to the right people. Having more connections won’t help your Network Impact, but having influential connections will.

Accuracy & Transparency
The subscores contribute to one overall score, the Klout Score. We’ve always been transparent about the various activities that could impact your Klout Score but we now have the power to share the specific actions that are helping or hurting your score. When your Klout Score changes you will be able to match it to a corresponding change in one of these subscores and understand why the change has occurred. If your Score goes up because more top influencers are acting upon your content, we will share that with you.

Influence is the Ability to Drive Action
The core premise behind our algorithms has always been that influence is the ability to drive action. We have tightened this concept even further in this release. You are not more influential because you tweet or use Facebook more, you are influential because you have an influential audience engaging with your content.

The Standard for Influence
With thousands of companies and millions of people leveraging the Klout Score, we take our role as the standard for measuring influence incredibly seriously. We are very early in what we view as a long journey. The team here at Klout is thrilled about the challenge ahead of us and are completely dedicated to creating the most accurate measurement of influence in the world. To that end, you can expect the way we measure influence to continue to evolve as behaviors change on the social web or as new networks like Google+ emerge. The majority of the time these changes will be incremental and invisible to most people, but this world changes fast and occasionally you can expect us to make significant changes like the one we are launching next week.

And of course, I know you want to know…

How will this affect my Score?
A majority of users will see their Scores stay the same or go up but some users will see a drop. Some of our Scores here at the Klout HQ will drop (including mine) — our goal is accuracy above all else. We believe our users will be pleased with the improvements we’ve made.

This is a project that’s been under development for over three months, and, in many ways, over the three years since Klout started. We appreciate your trust and support and we can’t wait to hear what you think. We will let you know when this new model goes live next week and will continue to work to provide the deepest and most accurate insights into your influence possible.

Posted in announcements, engineering, measuring influence | View Comments

#Kloutchat on Topics!

Tuesday, October 4th, 2011

I’m pleased to announce tomorrow’s #KloutChat will be on Topics! We’ll talk about +K, Topic Pages, topical influencers and much more.

#KloutChat on Topics

Wednesday, October 5th at 12PM PT / 3PM EST

Rough Agenda: (follow along with hashtag, #kloutchat to be a part of it)

  • Quick intro from Klout on our Topics and Topic Pages
  • Q1: What do you think of the topics on your Klout profile?
  • Q2: How often do you give +K to people? How do you choose who to give it to?
  • Q3: What do you think of Topic Pages? What’s your favorite part?
  • Q4: How do you think we could improve Topic Pages or better highlight them?
  • Q5: When you’re looking up others on Klout are their topics helpful to you?
  • Q6: If you don’t use Klout topics how else do you determine their expertise?
  • Q7: What else would you like to see from us regarding topics?
  • Open up to other questions.

This is subject to change, but we’re looking forward to a great convo! Remember, if your question if specific to your account it’s best to go to support.klout.com and get it answered there. Thanks!

Do you have any questions you want to make sure we address? Post them in the comments and we’ll try to add them to the agenda. #KloutChat is a regular event on the first Wednesday of every month. Note that we’re continuing to experiment with new times — let us know if this is good for you.

If you’re unfamiliar with Twitter chats, check out the article “What is a Twitter Tweet Chat?

Posted in measuring influence, social media | View Comments