“@kevinrose how do I get you to follow me back???” – anonymous self-proclaimed “social app guru”

This is Part 2 of a series that explores the science of Social Media Measurement. Let me preface this post by saying that this is a lighthearted post trying to come up with a simple measurement regarding a hugely successful social web service.

Previously, I explored the measurement of popularity, novelty, and attention on the very popular crowdsourcing news aggregation site Digg. My post was based on an arcane academic study involving the half-life of popularly “dugg” items. It turns out that stories frantically “dugg” by Digg members that make it to the coveted front page have a half-life of only 69 minutes. That’s a lot of work for a relatively short period of attention. Having that knowledge should prove useful to some marketers.

Why do we want to be able to measure social media? Why should we attempt to develop metrics? The ability to measure our efforts gives us valuable information to guide changes or course adjustments. It gives us baseline comparisons to measure against. It allows us to set measurable goals. It also helps us to make decisions when considering outside consulting help. In the social media space, there is a preponderance of self-styled “social media gurus” or “social app gurus” who try to trade off of non-existent influence. Like snake oil salesmen, they make grand claims about their reputation and expertise but their products or services are essentially worthless. Perhaps a few hard measurements could help marketers and advertisers identify the frauds from the really reputable experts.

This time around, I’d like to keep it a little lighter with a simple measurement I came up with to measure that amorphous quality called influence. This is a light-hearted attempt and nothing nearing the scientific exactitude I cited in my previous post. I consider influence as much more important than popularity, novelty, or attention. Indeed, influence implies popularity and attention. More specifically, I’m measuring influence on the hugely popular Web service known as Twitter. Twitter is a micro-blogging or mobile blogging service that essentially asks the question, “What are you doing right now?” When used correctly, it can be a helpful service for networking, sharing ideas, and staying abreast of buzz.

Twitter’s format is conducive to understanding and measuring influence because of its reciprocal structure of “follows” that makes for easy measurement. You can elect to “follow” other members of the Twitter service. Every time someone you are following “tweets” about something, you will get that update on your cell phone. People who elect to follow you are “followers.” Your followers get an update every time you twitter about what you’re doing or thinking.

Initially, most people “follow” their friends and family but eventually move to following other people on the Twitter network. Implicitly, people follow each other because they find each other interesting. I wouldn’t want to be following someone who is telling me and the whole world, “I’m going to the bathroom.”

There is a small group of highly influential members of Twitter that are so interesting and have such important thoughts to share that they quickly draw a whole army of “followers.” Their follower bases grow organically, naturally, and virally because they add a lot of value to the network and their followers. They don’t need to actively campaign for followers. Not surprisingly, their follower base is much larger than the number of people they follow.

There is also a larger group of sycophantic, self-branded “social media gurus” or “social app gurus” that have very little actual influence. We all know a few of these “leaches” and if they weren’t so spammy, they’d actually be mildly amusing. They are actively trying to get more followers. They spend a lot of time in self-promotion mode. They kiss your butt and play nice so that you might decide to follow them. They trade “follows” like high schoolers in a popularity contest. So instead of “I’ll vote you for best looking if you vote me for most popular,” they say “I’m following you so will you please follow me too?”

Not all “social media gurus” are frauds. However, you can spot the ones that are frauds when they try to build their follower base by asking truly influential Twitter members questions like this, “@kevinrose how do I get you to follow me back???” This is one case of a self-proclaimed “social app guru” asking Kevin Rose, the founder of Digg and a member of Twitter, to follow him. How inane is this? Please get a life.

So let’s get right into the Twitter Influence Ratio. It’s very simple really and very similar to the price to earnings or PE Ratio found in stock investing. In the financial PE Ratio – you get an idea of how much in earnings you getting for every dollar you pay for the stock. It’s a nice, convenient measure of how much value you’re getting or your bang for the buck.

Stock Price / Earnings = PE Ratio

EMC Corporation: 15.56 / 0.77 = 20.21

In the above example, EMC Corporation (EMC), a data storage company, saw its stock close at a price of $15.56 for the day. During the last twelve months, EMC earned $0.77 per share. Dividing $15.56 by 77 cents gets you a PE Ratio of about 20.21 – pretty simple right? Essentially, what the PE Ratio tells you is that for EMC Corporation stock, you are paying approximately $20.21 for every one dollar of earnings. Like I said, bang for your buck.

With the Twitter Influence Ratio, we’re going to try and get a read on someone’s true influence level. It stands to reason that if you are interesting, have neat thoughts, and add value to the network, people will naturally gravitate to you and “follow you.” Some of the most influential members of Twitter have many more followers than people they follow. So the Twitter Influence Ratio will attempt to express this relationship as;

Followers / Following = Twitter Influence Ratio

Example: 533 / 609 = 0.875

In the above example, one such self-branded “social app guru” has 533 followers and is following 609 others. This gives him a Twitter Influence Ratio of only 0.875 which means this person is not very influential. Intuitively, you ought to have more followers interested in what you have to say than the number of people you’re following. One might say that 533 followers is nothing to sneeze at. I agree, but the fact that this person has so many followers and is following so many more makes it highly probable that he is what is known as a “friend whore” or “follow whore.” Like the desperate high schooler, he’s just trading votes. Someone with a TI Ratio of less than 1 but is only following 30 others is probably not out there actively trading votes or follows. If I were looking for a consultant, I would run away from this guy and find someone more influential.

Let’s take a look at some folks who are truly influential. The aforementioned Kevin Rose, founder of Digg, is one of the most influential members of Twitter. As of this writing, he has a Twitter Influence Ratio of:

18,416 / 72 = 255.77

The Twitter Influence Ratio attempts to give you a sense of how influential someone is. In Kevin Rose’s case, for every one person he follows, he has just over 255 persons following him.

Justine Ezarik, or iJustine, a talented web designer is another influential member of Twitter:

12,652 / 1,047 = 12.084 Twitter Influence Ratio for iJustine

Of course, the TI Ratio doesn’t always work. If I claimed that it did always work, you could peg me as one of those phony, self-branded “social media gurus.” In the case of Robert Scoble, one of the most influential journalists and bloggers in the technology industry, he actually has a really low TI Ratio:

20,939 /21,243 = 0.985

Scoble’s a journalist so he has to follow as many people as possible to get the scoop. He is basically following as many people as he has followers. His Twitter Influence Ratio is almost a ratio of 1.

Well, there you go, the Twitter Influence Ratio is not a perfect measure of influence. But it does give you a sense of who is truly influential and who is just pretending. OK, this will probably be the last time you here me talk about the Twitter Influence Ratio. Tell that “social media guru” or “social app guru” you know to stop his Twitter spam before you “unfollow” him!

Update: I just found someone, Andreas Gohr, who wrote a script to tell you if someone on Twitter is likely to be a spammer. It’s based on the same principles as the Twitter Influence Ratio and is very well thought out. It helps classify users on Twitter in these categories: Newbie or Social Climber, Twitter Spammer, Twitter Caster, Notable, and Socially Healthy. Good stuff. My friend, the “social app guru” is definitely a Twitter Spammer.

Much attention has been given to the Web 2.0 generation of social networks and websites. Deservedly so, this next wave of Internet properties has quickly acquired humongous user bases, rich valuations, and cultural buzz worldwide. Venture capital investors are clamoring to fund the latest spin on crowdsourcing, content aggregation, social networking, micro-blogging, video sharing, and other different takes on social media. Fortune 2000 companies are all trying to figure out how to respond to this phenomenon; after all, they do need to be where customers congregate.

One of the big problems in this space is that advertisers and marketers are being duped by self-branded “social media gurus” and “social app gurus” who market their simplistic consulting and shoddy development services while demanding rich retainer fees. They promise “viral” adoption of inane and mundane “social networking apps” that lend little credibility to the corporate buyer. They promise “branding” efficiencies and opportunities by continually repeating their mantra of getting their clients “into the conversation.” But all good skeptical buyers should ask, “What does that conversation bring me?” More importantly, “How do I measure the effectiveness or ROI of anything that I am contemplating buying?”

At the root of the problem is the inability of the self-styled “gurus” to measure their effectiveness. Granted, social media and social networking are new fields and reliable metrics have not yet been developed. Nevertheless, progress is being made and this series will explore the emerging field of Social Media Measurement in order to bring some light and hope to what seems like a dark swamp of self-promoters looking to feed at the corporate trough.

Popularity, Novelty, and Attention

These three components are difficult to measure, but are critical to understanding how well a certain social media campaign is running. Any “guru” that advertises their expertise ought to be able to explain popularity and attention. Novelty is a little easier to define, we know what is new when we see it.

One of the great movements of Web 2.0 is the method of crowdsourcing. Crowdsourcing can be defined as “leveraging a community (this needs to be more thoughtfully defined) to gather intelligence or opinion.” It is what James Surowiecki explores in his book, The Wisdom of Crowds, and what Tapscott and Williams explore in their book, Wikinomics: How Mass Collaboration Changes Everything.

One of the great companies that has figured out a way to harness the power of crowdsourcing is Digg.com, a website that allows users to essentially vote or “digg” for interesting stories, articles, and blog posts. The items that receive the most votes rise to the top and garner even more attention. It is a useful website for finding out about what is important to others. Essentially, Digg is a good source of finding out about a particular story’s popularity, novelty, and attention.

Fang Wu and Bernardo Huberman, two researchers at Hewlett-Packard’s (HPQ) Palo Alto, California laboratory recently released a study exploring the mechanics of Digg. They essentially developed a mathematical model that describes how the popularity of a story decays over time. This decay algorithm is not unlike that which describes the half-life of radioactive material. Another victory for the Latticework Model of multidisciplinary scientific inquiry. At the core of their model is a function called a stretched exponential relaxation, which achieved the ability to measure the half-life of prominent Digg stories. Digg’s extremely popular front page, which has the power to generate vast amounts of Web traffic to stories that make the front page, features a half-life of 69 minutes! It takes a lot of work to get onto the front page of Digg but all that cumulative work results in a premium of just over 1 hour of premium attention.

Critique

I do have problems with this study and its conclusions. My issues aren’t with the conceptual idea of measuring popularity, novelty, and attention. I have caveats about their study design and assumptions. One of the well known secrets of Digg is that there are relatively select groups of Digg members who hold disproportionate influence in the community. Almost all the stories they “digg” or submit to the community habitually show up on the front page or home page. Everyone knows that there are ways to game the Digg engines. One could imagine loose “tribes” of Diggers that “help” each other digg stories in a reciprocal manner in order to get their stories to the top. This reciprocity has little to do with the novelty or newsworthiness of a story.

Another problem with the study design and assumptions is that a community is only as good or valid as its component parts. In other words, the demographic makeup of a community obviously is an overwhelming factor. Digg is a notoriously young and tech-savvy community. As such, it has its own built-in biases. Any study should take this into consideration. Most of the stories that make it to the front page are going to be technology-focused. And even within technology, the focus is primarily on Web 2.0 stories. One would find the community very hostile to Microsoft (MSFT) and other older, first generation technology companies. I think this is a secondary concern, but it should definitely be taken into consideration.

I would have liked to see Wu and Huberman explore the concept of influence. Those super-diggers who have an extraordinary ability to push their stories to the very top. In my next post, I will explore a very rudimentary measure of influence based on another very popular social media Web service.

Implications

What are the implications of this story? Perhaps a corporation contemplating a viral campaign on Digg to try and push awareness of their story could measure the costs of such an effort. Would the ability to identify the super-diggers who might be willing to trade influence for monetary compensation be a worthwhile endeavor? Note that the community would frown upon a paid campaign to push awareness. Are the risks of negative attention worth the benefits of getting a chosen story to the top? Can 69 measly minutes of premium attention be worth the costs?

Again, I issue a warning to corporate buyers. Be very skeptical of self-branded “social media gurus” and “social app gurus” – they likely are not able to quantify their expertise and will be unfortunate resource wasters. Believe me, the talent in this cottage industry is very weak but great at self-promotion. The emergence of the new Social Science will alleviate this problem by helping advertisers, marketers, and strategists identify and measure ROI and performance. New metrics are arising everyday.