On Monday, I asked Twitter users to complete a very short survey. They answered questions about how many followers they had, and how many clicks a link received when they shared it using a service like Bit.ly.
I asked because I wanted to get a sense of the clickthrough rate for links shared on Twitter. I received about 45 responses. Not enough to qualify as statistically significant, but a good start nonetheless.
To get the average clickthrough rate, I just divide the total number of clicks reported (1656) by the total number of followers reported (46973). That gives us a rate of 3.5%. This means that, on average, a link shared on Twitter will be clicked by 3.5% of your followers.
As you might expect, there’s a relationship between the total number of followers and the clickthrough rate. For those with less than 800 followers, the clickthrough rate is 8.6%. For those with more than 800 followers, the rate falls to just 2.5%. This make sense. If you have a small group of followers, you probably have a closer relationship, on average, with each of them. They’re likelier to click links that you share.
Thanks to those who participated in the survey. If you’d still like to contribute, take 18 seconds and do so.
I’m running a quick little survey about clickthroughs on Twitter. It applies to you if you use a link shortener like Bit.ly that enables you to track the number of clicks a link receives when you post it on Twitter (hmm…mangled sentence there, but you get the idea). It’s all of two questions, and should take you, like, 14 seconds:
On this day, two years ago, I generatedthis chart using tweetVolume. It shows how frequently each of these words occurs in Twitter conversations:
Here’s what that same chart looks like today:
It’s somewhat biased by the current NHL playoffs, but it’s interesting to see how the term ‘php’ still dwarfs the other terms. Not particularly scientific, but does it indicate that Twitter’s audience still skews geeky? Or maybe it just shows that the Twitter power users are a nerdy bunch?
On this site, I prefix client-related stuff with ‘Client Plug’, or indicate something similar in the opening paragraph. I also try to make it clear that I was invited to an event for free or received something for review. This whole blogger disclosure discussion became particularly important with the advent of sponsored posts and links.
But what about Twitter? There are plenty of reasons to like the 140 character limit, but it hardly encourages transparency. There’s rarely enough space to disclose one’s interests. On more than one occasion, I’ve witnessed people tweeting about projects with which they’re affiliated without making that association obvious. The most common scenario is when Person A tweets about something, and Person B replies with “wow, that’s a fantastic thing, good luck with that thing” when Person A and Person B in fact work together.
I’m certainly guilty of retweeting client projects without disclosure. Here’s an example from today:
As I mentioned, we’re working with ActiveState on a new project. And you’d see their name on our client page on our company site. Yet a casual Twitter follower might have no idea of the formal, financial connection between myself and ActiveState.
Unicode to the Rescue?
A common solution here might be a hash tag, such as #clientplug or #disclosurepending or something. That would do, I guess, though it already takes up a fair number of characters. If you’re interesting in seeing a message spread, then space is already at a premium.
Here’s a silly idea. What if there were a generally accepted unicode character that you could add to tweets to imply that you had a personal stake in the message? A kind of disclosure shorthand that people could follow up on if they had questions. Maybe it’s â˜Â, as in “I have strong connections here”? Or maybe â˜Â, as in “I’m promoting this organization”? Or maybe just ♟, as in “I’m a tiny pawn in this giant corporation”?
What do you think? Does disclosure on Twitter even matter?
Skimming recent new followers this afternoon, I discovered that Ms. Denise Savoie, Member of Parliament for Victoria, is newly on Twitter. I started wondering which Canadian MPs are on Twitter?
At South by Southwest Interactive last week, marketer Peter Shankman said “if you say you’re a social media expert, I’m going to check how many Twitter followers you have.” This is about as useful a metric as saying “if you say you’re a professional hockey player, I’m going to count how many hockey sticks you have.” It tells only a tiny fraction of the whole story.
Shankman’s comment got me wondering: how would Twitter be different if the service didn’t publish statistics about who you’re following and who’s following you? Because these numbers are public, we’re experiencing a kind of follower arms race, where heedless reciprocal following has become the norm and popularity and leader-board sites are de rigeur (I’m still working on getting my douchebag index into the nineties). One’s list of followers has become, for better or for worse, the new unweeded blogroll–messy, too long and polluted with hastily-exchanged links. This shouldn’t be a surprise: from box office revenue to Technorati ranking, if we can count it publicly, we will.
Severalpeoplehavebegun wondering about how our burgeoning networks will scale, and how its users will deal with the growing amount of signal.
If our Twitter numbers were private, wouldn’t we be more selective about who we followed? Wouldn’t we focus on forming a conversational network based on the quality of people we followed, instead of the number of people who followed us? Wouldn’t we emphasize meaningful discussion over the “top seller” mentality that seems to pervade the tool?
Grade Eight Gym Class
It took me years to start obsessing about the web stats for this site. I eventually learned to follow my friend Dave Olson’s advice to “fuck stats, make art”. Now I’m facing a similar grindstone with Twitter. The bloggy, social media world has always been a bit too much like high school. I thought we’d begun to grow out of that immaturity, but the tyranny of Twitter stats puts us right back in the sweaty locker room after grade eight gym class.
I was talking to my friend John Keyes about how I could reduce my compulsion to watch my Twitter stats. He was sympathetic, and whipped up a Greasemonkey script that simply hides the follower numbers from the tool’s web interface. It’s only as effective a mind hack as, say, setting your watch five minutes fast. However, it’s a little reminder to chill out and use Twitter the way I want to, instead of how the popularity-obsessed web demands that I do.
UPDATE: British writer and comedian Dave Gorman has a great post that touches on a similar topic:
But yesterday I had two people contact me to tell me that I was rude for not following them. How not-following someone can be rude is quite beyond me. So I asked. And their point was that they were following me and that it was therefore only polite for me to follow them back because unless I did that I wasn’t being interactive.
Which seems to me to be a false definition of what interactivity really is. In what way would clicking a button to say I was following someone be actually interacting with them? At the moment I follow between 200 and 300 people. When I log on I normally find there are between 10 and 20 posts for me to look at from the last 5 minutes of activity. But I’m followed by over 20,000 people. If I followed all of them, there would be a hundred times as many recent posts to review. There would be no way of me actually reading – or even meaningfully scanning – 1000 to 2000 posts every 5 minutes.
I especially like his conclusion:
The difference between following someone and replying to them is the difference between stopping to chat with someone in the street or giving them a badge declaring that you know them. One is actual interaction. The other is just something you can show your friends.
I tried to write this post a couple of times, but faltered. So, I figured I’d try to articulate myself using video. The result, I’m afraid, is really no better. Remember–that’s four minutes of your life you can’t have back.
Building on what I ramble about in the video, consider the example of somebody receiving 1000 tweets a day. Let’s imagine that they actually read 200 of those. The other 800 just float by in the endless Twitter river while they’re working, interacting with other humans and so forth.
If each person in their Twitter network posts 10 times per day, then, on average, 2 out of 10 of each person’s tweets get seen.
Now imagine that the size of the average network doubles, to 200. That means 2000 tweets a day. The user still only sees an average of 200, so only 1 out of 10 tweets get seen.
Everything increases but our attention bandwidth. Is there some kind of threshold where the river o’ Twitter becomes too diluted? If the average follower count continues to go up, will we someday rely almost exclusively on DMs and @ messages? Or, as I speculate about in the video, will we just get better at filtering and personalization?
Before Christmas, while working on the Save the Great Bear campaign, I had an idea for a Twitter-powered site that could, if successful, be an effective engine for spreading news and calls to action for social change causes. I talked with sundry people about it, including Rochelle and Geoff (they made Twemes), Joe, Boris, James and so forth. Everybody seemed to think it was a good idea. So I submitted it to this contest on Changemakers.net (please consider rating my idea on the site–no registration required). Even if nothing comes of that (and the odds are against me), I’m hoping to build it.
That’s it in a nutshell. After the jump I’ll excerpt some text from my Changemakers.net submission, which explains and expands in text what the video summarizes. Just to clarify, there’s nothing at 10000Tweets.com–the thing hasn’t been built yet.
