Gaming the System – Why Follower Counts Don’t Represent Influence
In December of 2009, I wrote about Misleading Indicators – Followers and Friends after seeing a tweet from Jeremiah Owyang of Altimeter Group. In that post, I explained why follower or friend counts do not represent a metric of influence nor should they be utilized as a relevant metric of importance. After several good comments as well as several emails to Impact Interactions giving me grief for disputing one of social media’s closely held beliefs, I decided to run an experiment on gaming the system.
My basic premise was that these counts are a false statistic which like placing value on “hits” in web metrics analysis leads you to focus on the wrong metric of your activities. Want to increase “hits”? Add more photos, widgets, content blocks, etc. to each of your pages. Each one adds a hit each time the page is opened. You can make the hit count be anything you want simply by adding more items to each page. In 2000, most people didn’t understand that aspect of the measurement so they used “hits” as a proxy for visits or even for influence of their content and site. We still have companies that talk about “hits” when they approach us about measurement. It’s a lasting issue that has thrown a lot of folks away from the important issues in measurement.
There have been several blog posts written about how to game Twitter to gain followers in order to look more important than your competition. One of my absolute favorites is from Chris Cree of Success Creeations. His blog entry “How to Game Twitter to Add Thousands of Followers Every Day” should be mandatory reading for all social media marketing professionals. It spells out how you can game Twitter, but also why that is such a bad idea. So with that advice in hand, we set up a little experiment using free tools to game the system.
The Experiment Begins
Using a tool we won’t name here, we were able to almost double our number of followers for our @impactinteract twitter account in five days. Granted we were working from a small base, but the results show how easy it was to pull off. So let’s go to our experiment.
We started last week with 143 followers who found us either from our website, our efforts on LinkedIn, our speaking engagements, or organically from our tweets. We were following 43 members who were mostly our competitors. On Monday, I signed up for a free demo of one of the many tools which advertise that they can add followers quickly. By using the key words of “Social Media” and “Online Community” the tool returned over 700 accounts on Twitter that had potential for us as followers. These accounts had either tweeted the key words “Social Media” or “Online Community” in the past ten days. Sounds good so far right?
The tool then allowed us to follow the accounts in order to grow our followers by getting their auto-follower to reciprocate. The demo of the tool we choose allowed us to generate up to 250 new followers before we would have to buy a license. So we started the process using the tool of following 250 accounts. It was fast and painless. In the fifteen minutes it took to follow these accounts, we were able to work on other activities. Once the 250 follows had been accomplished, we waited about a day and then unfollowed any account that didn’t auto-follow us. Over the next several days, we repeated the steps. Here is the table of our activities:
To keep everyone who autofollowed us aware of what we were doing, we tweeted a message several times during the experiment that stated:
We are testing a few of the tools that advertise that they can build your follower base for an upcoming blog #socialmedia #Twittermarketing.
The idea was that if the new followers actually read our tweets they would also know what we were doing. That way they could unfollow us as quickly as they auto-followed us. Incredibly only 9 new followers over the course of the week unfollowed us. None sent a direct message about what we were doing. So in a little over a business week, we came close to doubling our followers. Total time including the time to download and set up the tool was about 2 hours total.
“Ah ha, the tool worked!” you might be saying to yourself. But did it really add followers for our corporate Twitter account who might spread our message and help us grow? Let’s take a look and find out if our tweets on social media and online community news and trends, as well as our company news is really relevant to our new found followers.
Of the 136 new followers, 14 (10%) sent the same auto-messages to me about making money on my tweets:
MAKING MONEY for your Tweets? I am. Making 20 daily on autopilot. Make money too – TODAY! http://bit.ly/xxxxxx Thanks for following
Another 9 (6.7%) sent an auto-message inviting us to join their multi-level marketing scheme or affiliate marketing network:
Thank your for following me at http://bit.ly/xxxxxx. We’re looking for affiliate marketers to help us. Do you know any?
Welcome to AffilBits! Want to know how to get thousands of targeted Twitter followers and earn a 50% affiliate commission at the same time?
Two follows were from famous and semi-famous people: Emma Watson of Harry Potter fame and a porn star.
12 (9%) new followers were from two unique members who used multiple accounts, but the same photo.
So out of the 136 new followers, we found 37 (27%) were not, nor would they ever be interested in Impact Interactions.
Influence scoring of our new followers shows the truth in the fallacy of follower counts. We used a scale of 1 to 5 to rank our new followers in regards to our ability to be influential with them or in their networks. A score of 1 means Impact Interactions is not potentially influential at all, 2 means probably not potentially influential, 3 means neither potentially influential nor not influential , 4 means somewhat potentially influential, and 5 means Impact Interaction is potentially influential. (And yes, we understand that this is not scientific because we are making the judgement. But how many people on Twitter really analyze their follower base on an individual level?)
Our influence score would be 1 with the group of 37 detailed above.
But what of the other 99?
We reviewed their tweets over the past ten days to see if these would really be good followers for us or not. What we found was 65 were simply folks who had retweeted someone else’s message about a social media topic. They were neither working for companies involved in social media or online communities nor were they particularly interested in the topics based upon analysis of their tweets. In fact several of the members were serial retweeters. We went back through several weeks of tweets and never found a single tweet that they created. So our influence score for these twitterers would be a 2.
There were 7 new followers who are in the search engine optimization industry, another 20 who are potential competitors or individual consultants trying to find work in the social media industry. The influence score for these followers would be 3.
The remaining seven new followers were blog publishers creating newsletter style blogs of others’ content around social media and online communities. They were linked to content aggregation sites rather than competitors. As these sites could potentially help us to influence their readers, we gave them an influence score of 4.
There were no members of our target audience of corporate social media or online community management staff amongst our new followers.
The weighted influence score for our new members was better than we expected at 2.02 (meaning Impact Interactions is probably not potentially influential to this new group of followers).
So, does our newly increased follower count mean that we’re more influential in the social media and online community world? No, it does not. You shouldn’t be impressed with the number of your Twitter followers either. With the set of tools available today, you too can gain thousands of new followers in days. But those followers won’t buy into your view of the world or your brand. In many cases those counts have been culled from the Twitter Borg, not from an audience that cares.
Organic growth of your audience builds an audience that actually is interested in your message or company. Use your content, flair for creativity, and on-target messaging to grow your followers. Include your Twitter account information (@ImpactInteract) in your email and other outgoing communications. Your influence will be stronger, even if your follower counts are smaller. Bigger isn’t always better, but don’t buy into the myth that more followers equals more influence. If we don’t put an end to this measurement idea, we will be having the same discussion in five years that we do with “hit counts” today, more than ten years after it first came up.
To all of you who began following us during the experiment, thank you for taking part. If you wish to unfollow us, we’ll understand.
Mike Rowland, President
Back to the blog
This entry was posted on Tuesday, March 2nd, 2010 at 6:00 am and is filed under Best Practices, Measurement & Reporting, Social Media Trends. You can leave a response, or trackback from your own site.
9 Responses to “Gaming the System – Why Follower Counts Don’t Represent Influence”
MRowland says: Hi Alan,
Thanks for the comments here as well as on the LinkedIn Group discussion.
The issue of how much you can read in a day comes into play as mention. The use of Twitter Lists can help control this, but this tactic also reinforces the notion that followers are a useless metric. That’s because you may have the 100,000 followers but you are only paying attention to a small number rendering the others into the category of background noise. If you don’t know them, you cannot influence them.
As for me, I’d rather have that one Mercedes…
MRowland says: As a follow up to this post, we are still gaining followers from the efforts last week. Seems like some of the folks chosen to follow don’t actively use Twitter day to day. We’re also still getting a lot of followers who send a DM about making money on Twitter. We’ll be culling our list of followers shortly to get back to those who we can be influenced by rather than those who auto-follow without reviewing who they are about to follow. Some people just don’t get it.
That is an excellent experiment, Mike. Obviously I agree that adding followers for the sake of bigger follower counts is worse than pointless. I’ve held that position even longer than the year it’s been since I wrote that post.
But it’s nice to see that lack of value quantified in a meaningful way.
I figure that, from a business perspective, the focus of social media should be influence. You do need an audience to be influential and a bigger audience is often a measure of influence. But with social media, and especially on Twitter, influence is better measured by interactions such as replies, retweets and link click-throughs more than by totally follower counts.
MRowland says: Hi Chris,
Thanks for the comment (and the retweet). I agree that the focus of social media should be influence. We always begin working with our clients on the objectives for the project. We then get tactical, but we never focus on simple raw numbers only on finding an audience that will engage with the client’s message/issues.
I really liked your initial post on the topic. I’ve been pleased to see more blog posts that are similar to this one emerge over the last couple of days. If we all continue to push the idea of quality of audience over total audience counts, clients and companies will find their results much more satisfying from an economic value point of view.
Thanks again.
Mike
[...] of the Social Media Specialist. But is it really? We’d say it isn’t. Read our Gaming the Twitter System: Why Follower Counts Don’t Represent Influence to learn [...]
This topic is getting some interesting comments from the Consutancy Group on LI. check it out.http://www.linkedin.com/groupItem?view=&gid=40103&type=member&item=29139961&qid=29ba67a3-1cc1-45f0-a68a-9d54eaf4e8a4&goback=%2Egmp_40103
[...] while your audience has the advantage of learning more. B2B social media requires measurement. Friends, followers, group members, etc. don’t mean anything until they are engaged in a conversation or contact process with your firm. Why? Because you cannot [...]
[...] and products. (To learn more about why followers and fan counts are not a measure of influence, read this.) Audience building takes compelling content that excites and interests your potential customer. [...]
The basic problem is confusing numbers with influence and quantity with quality.
If I have 100,000 followers, I can reasonably follow only a few on a regular basis. I assume the same is true for most who follow me. If so, the “impact” of a mass following is illusory because few of them will ever see my tweets.
Influence also derives from retweets. In the above scenario the number of RTs my tweet generates will be minimal.
Finally, gaming the system assumes all followers are equal, which they are not. A Jane Doe RT is not equivalent to an industry heavyweight RT. In short, what is more valuable: 10 Yugos? Or one Mercedes?