Klout is the self-proclaimed "standard for influence" and even claims that "when we're measuring your influence there's no room for error." Anyone who is active in social media (facebook, twitter, etc.) can sign up to have their influence measured by klout. According to Klout, your score is based on three factors:
- True Reach: How many people you influence
- Amplification: How much you influence them
- Network Impact: The influence of your network
Something happened yesterday (October 26, 2011). As promised, Klout made some changes to its algorithm and peoples' scores changed. This has happened in the past, but yesterday's change seemed more radical than others, with scores dropping by more than 10 points in many cases and leaving people asking why? What changed?
I have a theory: AMPLIFICATION
I believe that the way they measure true reach and network impact remained fairly consistent. However, the way that they measure amplification changed significantly.
In the page that describes the klout score, amplification is described as:
Your Amplification is how much you influence people. When you post a message, how many people respond to it or spread it further? If people often act upon your content you have a high Amplification score.
I don't think they changed this description yesterday, because this is how I remember amplification having been described in the past. However, if you look at your profile page and hover over the "?" next to the word "Amplification" it is described as:
Amplification is how much you influence people. It indicates how likely your audience is to respond and how close you are to your entire network.
Do you see the difference? They changed it from a "how many" measurement (i.e. how many people reply, retweet, comment, thumbs up, etc.) to a "how likely" and "how close to entire network" measurement. So, someone who has a small, but very engaged audience is likely to have a higher amplification score than someone who has a large and partially passive audience.
There are hints at this in Klout's description of the changes too. They say:
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.
So, if you have 5 followers, you tweet something, and they all reply and retweet that, you are going to have a very high amplification score. However, if you have 100,000 followers, and 300 of them reply and retweet, you are going to have a low amplification score.
Let's look at an example
My friend Maranda (@hotelqueen) has a klout score of 60 (hers did drop, previously it was 68). I (@phdinparenting) have a klout score of 58 (I believe mine was previously 72, but I'm not 100% sure).
Now, let's compare some stats for Maranda and I:
@hotelqueen (Klout Score = 60)
- Followers: 1,831
- Following: 722
- Klout True Reach: 4K
- Klout Amplification: 54
- Klout Network Impact: 33
@phdinparenting (Klout Score = 58)
- Followers: 61,552
- Following: 2,840
- Klout True Reach: 8K
- Klout Amplification: 11
- Klout Network Impact: 42
Looking at these stats, you can see that Maranda's amplification score almost five times as high as mine is, even though my stats are higher across the board on the other factors.
Why is that?
Maranda may have a relatively small number of followers, but they are incredibly engaged with her and her content. I have a larger number of followers and I may have a higher number of re-tweets and replies overall than Maranda does, but percentage-wise a larger portion of her audience is engaging with her than my audience with me.
Ultimately, what it means is that being followed by a lot of people, even those super-popular and super-connected people, isn't going to get you anywhere unless they are actually replying or retweeting your content. In fact, you are going to be penalized for the people who follow you and don't engage with you. So instead of people getting pissed off at Chris Brogan for his massive unfollow of EVERYONE, people should be thanking Chris for boosting their klout score.
The conclusion...
You should follow Maranda. She's awesome and the people who follow her can obviously attest to that. But don't follow her if you're just going to be dead wood and not engage with her. We don't want her klout score to drop. :)
And if I really cared about my klout score, I would have to spend the good part of a week blocking thousands of followers who don't ever engage with my content because having them passively follow me obviously makes me less influential (because, you know, there is "no room for error" in klout's algorithm).
UPDATE -- CONFIRMATION
Today (November 22, 2011), Klout added some additional information to the klout profiles. It now lists the 90-day activity for each user. This essentially confirmed my thoughts about how the amplification score now works. Here is the information on my klout score/activity (@phdinparenting) and Maranda's klout score/activity (@hotelqueen).
If you look at the sum of my retweets and mentions (7.8K) as a percentage of my followers (62K), you get 16%. If you look at the sum of Maranda's retweets and mentions (1.6K) as a percentage of her followers (1.9K), you get 85%. That is pretty close to our amplification scores, of 10 (me) and 72 (Maranda). In both cases, our amplification scores are a bit lower than the retweets/mentions as a percentage of followers. My guess is that klout only counts each person once, so if 3 of my replies were from Maranda, then only one of those would count towards my amplification score.
I checked this with a few other people and the results vary. But that probably depends on what percentage of a person's replies and RTs are coming from the same people versus a wider spectrum of their network.
I continue to be intrigued, if not impressed, by the klout algorithm. I'm a math geek at heart, so I can't help myself.
