You may know Ben Folds from the ever impressive band Ben Folds Five (but if all you’ve heard is the song ‘Brick’ and you think you know what they’re about, go listen to this or this or this and come back when you’re done). And William Shatner? ‘Nuff said.
Well I’ve been a fan of BFF and Ben’s solo music since college (don’t ask when that was). In fact, I confess - I once listened to “The Last Polka” 30+ times in a row on cassette tape in my dorm room (yep, had to keep rewinding the tape). No that’s not a joke and no mind altering substances were involved.
The Myth Of Raw Follower Count
Back to the point. It sounds so good on the surface. Five hundred thousand followers. Damn! Right?! I was surprised however, to hear Ben say he didn’t feel his tweeting was helping significantly to sell albums. Check out the clip, which should begin at 28:30;
The exact quote from Ben in the interview;
“Because I tweet something it doesn’t mean many people saw, it just means people could see it if they checked it. Even though I’ve been tweeting my ass off lately, I’m still getting people sending me tweets that go ‘What?! How did I miss this new Ben Folds Five record?’. I don’t know, cause I said it over and over again but this is the way it works, so we’re figuring that out. It’s a good theory, like, wow dude’s got half a million Twitter followers, he should be able to sell some records. But we’re finding that turning that energy in the direction of getting the word out there is actually very difficult.”
500,000 followers on Twitter… doesn’t help spread the word about a new album? Or at least not to the degree you’d expect?
Well Ben I have the answer to this mystery – some data digging through FollowerWonk. I think we’ll all find the data a bit enlightening, and perhaps even surprising.
What is FollowerWonk you ask? Analytics for Twitter! Let’s look at why out of 500,000+ followers, more people aren’t seeing your tweets.
First, Run Ben’s Followers Report
Before we analyze, we have to run the report. So we’re just going to go to FollowerWonk.com and run an “analyze followers” report. Enter “benfolds” – and then wait about 10-15 minutes for it to run (since he’s such a popular guy – that’s a fair amount of data).
BTW – you can click on all photo links for bigger view.
Anyhow, since he has 500,000+ followers – FollowerWonk will run a sample size of about 100,000.
So now that we’ve run the report, we can dive in and analyze his followers and see what’s up. There seem to be five plausible reasons why more people don’t see Ben’s tweets.
1) 52% Of Followers Haven’t Tweeted In Over 30 Days
So Ben is very correct in his intuition about Twitter. If half of his followers aren’t even on Twitter more than once a month, odds are very slim they’ll see something, especially if he just tweets it a few times.
2) 78% Of Followers Have Tweeted 499 Times Or Less
Second indicator of low activity of his followers. Most people I see that are pretty active on Twitter, have at least tweeted over 500 times. Five hundred or less tweets, to me, indicates someone who is on Twitter very infrequently, or used Twitter for a while and hasn’t been back in months or years.
3) Followers Time Zones Are The Opposite Of Ben’s
Far as I recall, Ben lives in Australia (note that if I am wrong about this, this is the least concerning of metrics, BTW). And although when touring, he’ll be on US time zones, check out the most active times of his followers. 12pm – 12am NYC time. This would mean Ben needs to do most of his tweeting from 2am to 2pm Australia time for the most people to see it in their time zone. The dip is huge, but maybe he’s up past 2am or tweeting before 2pm.
UPDATE: I stand corrected, that Ben now lives in Nashville. Thanks Matt for the comment below.
You can see further in this map (which is just a small sample size of about 5,000 followers) most followers are in the US. Tweets during the day in Australia won’t get in front of many of those people sleeping in the US.
4) ‘Follow Back’ In Profile Possibly Indicates Low Quality Accounts
This is a word cloud of two word phrases from Ben’s followers bios. You can see “follow back” is the number one most used phrase. In my experience, those that put this in their Twitter profile are not likely to be high quality users. They are sometimes accounts that are just looking to try to get a bunch of followers.
5) 25% Of Ben’s Followers Follow Too Many People
I dunno about you guys. But following only about 100 people makes it hard to keep up with my stream. I can’t imagine 300 or 500 or 1,000. But approximately 25% of Ben’s followers follow over 1,000 people. I’m sure they miss a tweet or two Another likely reason why Ben’s tweets wouldn’t get noticed as much as you’d imagine.
500,000 Followers Is Not His ACTIVE Follower Count
Ben may have 500,000+ “followers”. But how many of those are actually active. How many people might actually see his tweets? When you add all five of the above factors together;
- only 48% of followers on Twitter in last 30 days
- only 22% of followers have tweeted more than 500 times
- most followers are opposite Australia time zone (if that’s where he still lives and tweets from often)
- bio word cloud indicate some low quality profiles
- 25% of followers following too many people.
I’m sure there’s a more robust way to estimate possible impression percentages. This is a REALLY rough estimate. But…
- Ben probably has more like 50,000-75,000 active followers
- Roughly 25% of people are following too many people to see every tweet.
- 75% or more are in US time zones…
Which means active follower count is more like….
- If you tweet on US time zones, about 30,000-45,000 people have an OK chance of seeing your tweet
- If you tweet on Australia time zones, about 10,000-15,000 people have an OK chance at seeing your tweet
Again. Totally not scientific with that estimate. I’m sure there’s a decent margin of error. But you get the rough idea. Raw follower count is a terrible way of estimating impressions or “klout”.
And I hope Ben knows, I do NOT intend to make this sound critical or negative in any way Cause none of this even questions the devotion of his fans, or the quality of his tweets. I’m assuming his fans love him, and they want to watch what he tweets.
This data goes to show – even with raving fans, and entertaining tweets – 500,000 followers is not 500,000 active followers. The two are very different.
Click Through Rate
Let’s look at one more metric – clicks. Ahhh… CLICKS! I’m glad Ben used a Bitly link on his very first tweet about the new album happening.
This tweet went out on Wednesday July 18th at 11:28am NYC time. That’s actually a great time for a tweet, Mr Ben! It got Re-Tweeted 226 times.
How did it do for clicks?
It got 7,177 clicks – the first 6,000 came from the day he tweeted it. That’s about a 1.4% click through rate. Assuming 500,000 people saw it. But that’s the catch. After the FollowerWonk data from above, I’d be willing to bet maybe 50,000 or less people actually saw that tweet.
But if you’re using active followers – you’d get a pretty healthy click through rate of about 14%. THAT is not bad at all. So for the amount of active followers, he’s moving clicks pretty well – and 14% makes a bit more sense. I don’t know what album sales are – so the final conversion metric is not handy, unfortunately.
The Best Imitation Of Yourself
So I have one tip for Ben. It’s a simple one, but for someone with even as many active followers, it could go a long way. You gotta do the Best Imitation Of Yourself and tweet more than once.
When there’s a really important tweet – or you’re promoting something and want clicks - tweet it three times within 24 hours.
Space your tweets about 6-7 hours apart. Load them into Buffer if you have to. Or Hootsuite. They’re both free – they function differently but you can schedule your tweets. Three tweets spaced about 6-7 hours apart will hit each area of the globe, and you could quite possibly double the amount of clicks you’re getting.
A few extra tweets could turn 7,000 clicks into 14,000 clicks, easy.
Moving Beyond “Klout” and Raw Follower Count
SEO’s – if there’s one thing even this rough study shows us: raw follower count is a terrible metric. So worthless it’s misleading.
BUT – I think FollowerWonk and the world of Twitter and social media metrics is extremely exciting and promising. We’re starting to move beyond the unreliable “metrics” of Klout and the even less reliable idea that number of followers can prove influence. With tools like FollowerWonk and True Social Metrics I think we’re finally stepping into the second generation of analytics for social media.
It’s SO new right now. Think about the calculations we have to do if we just wanted to estimate how many people saw a tweet. It’s crazy. Tools like TweetReach and CrowdBooster (which I like and use) measure “reach” but all this is, is total followers. I’d love to see a day when somehow actual impressions – actual eyeballs on a tweet is a standard metric. I think we’d all be a little surprised at who really has the ”klout” on Twitter.
I’d be really curious to hear anyone’s comments below. Including yours, Ben, if you’re reading