Meet TwitShowdown

This post was generated by Dashter

Over the past few weeks, I’ve really been trying to pay attention to the pro’s and con’s of Twitter. Part of that’s due to the fact that I shut down my Facebook account, in an attempt to see what the non-Facebook Internet still felt like.

Kelly and I had an interesting back-and-forth based partly around the idea that we follow people at odd intervals. Sure, it’s easy to add the first 20-100 people you follow on Twitter because you know them in real life. But then you start following people you’ve never met. And even the people who we know (or know of) don’t always tweet how we expect them to.

So this idea started forming – the notion of giving a way to easily swoop in to your account for the express purpose of unfollowing people. It wasn’t a “Dashter” feature – this was a unique little activity that needed a unique place to go.

Lose the Losers is right!

The Solution: twitShowdown (you can go visit it here: http://twitshowdown.com)

So this is the end result: A way to quickly unfollow people. But there’s a method to the madness, and I want to share (geekily) my thoughts on Twitter signal vs. noise, and the nature of social media attention span.

Social Media Attention Span

The real issue with Twitter is that it doesn’t have an “edgerank” filter like Facebook does. This is a good and a bad thing. The problem with Facebook is that they’re making judgements based on what they think you want to see – which leaves many of your updates off the radar of the people you’re friends with. Twitter on the other hand is virtually unfiltered. It’s a pure stream of consciousness – with one notable exception. That exception is by placing an @mention at the beginning of a tweet, it will only show up in the stream of someone who follows both people. Otherwise, it’s considered a conversation thread that is filtered out for you. That’s effectively the only filter on your Twitter stream. Everything else gets piped in.

I’m a big fan of big data. But humans aren’t really well built for big data – we depend on filters to provide us with relevant or meaningful subsets of data. Twitter is beautiful because it’s a big giant data pile, but it also costs you attention span to sort & sift.

Let’s do some back-of-the-napkin math on this. Let’s say that you have 30 minutes per day to interact with Twitter.

The average reading speed of a regular person is 410 Words per Minute.

In that 30 minutes of “Twitter time” you can read [ 410 x 30 = ] 12,300 words.

The “average” tweet consists of about 90 characters. Considering abbreviations, spacing, and the desire for brevity, let’s say each word is 6 chars long, yielding a word-count “per tweet” of 15 words per tweet.

So, assuming that you can read 12,300 words in your 30 minutes of Twitter time, and each tweet is 15 words long, your 30 minutes yields about [12,300 / 15 = ] 820 Tweets per 30 minutes. That’s around ~27 tweets per minute, or 1,640 tweets per hour, give or take.

Alright, so that’s how much you can read and comprehend. So what about your attention span?

On Twitter, it’s really easy to follow someone. But you don’t always know what you’re actually getting when you click that “follow” button. That’s one of the reasons we built Archetwypes in to Dashter – to give you a fair warning about someone’s tweet style and let you decide whether or not their methodology is aligned with your preferences and values. But what about people you already follow?

Let’s break it down this way. Let’s say you follow 100 people on Twitter.

Let’s further assume each person on Twitter tweets 3 times per day.

That means that you should see about 300 tweets per day. That’s easily read-able within your 30-minute daily “tweet-span.” You can (in theory) be up to date on everything in your Twitter social network every day, and still have time for out-bound engagement & conversation.

But what happens when you follow 500 people? Now you’re talking about 1,500 tweets per day in your stream – meaning you’re not reading almost 50% of the tweets you’re receiving. What about following 2,000 people? Now you are being exposed to roughly 6,000 tweets per day, but you’re only able to read 820, meaning you’re only getting about ~14% of the total signal.

And there is huge variation in tweet yield as well. Consider this screen grab from TwitShowdown (click to see full size):

Tweets per day ranges from 0.15, 0.41, 3.12, 6.17, 13.81, 17.42. The mean is 6.85 tweets per day (TPD) while the median is 4.65 TPD. Notice those heavy tweeters pull the mean drastically towards the high end. I’ll pull another random sample (not shown): 0, 0.57, 0.60, 0.68, 5.83, 11.34 [mean= 3.17; median= 0.64].

So why TwitShowdown?

Alright, so that explains the theoretical problem. You probably follow too many people on Twitter. In theory, you should be following about 10 people for every 1 minute you plan to spend on Twitter per day – in order to stay abreast of all the happenings. But we (as social creatures) don’t do that at all. We follow loads of people. And Twitter (rightfully) has made it super-easy to follow new people. It’s how they build their platform. But un-following isn’t quite as easy. We feel a pang of regret at the thought of missing someone’s tweets. There’s an economic / psychology term for what I’m thinking of, but I can’t remember it just this second. But effectively we convince ourselves that we made a correct decision by following someone and justify that decision in an ongoing fashion by not un-following them. Even if that person spews an supply of useless or un-entertaining tweets, we’re don’t exert the effort to un-follow because we believe we’ll be worse-0ff by removing them from our feed on the off chance that a future tweet will be well worth our while.

So, that’s why TwitShowdown looks the way it does.

It’s pretty simple. You log in with your twitter account, and you’re shown 6 random people that you follow. Once you drop someone, it’ll refresh and give you a new batch of people to choose from. The only real data you’re provided is their avatar, their name & screen name, their Tweets per Day, and whether or not they follow you as well.

The avatar & screen name are there because you’ll probably instantly associate their name & face. If they’re someone you know, or someone you remember being interesting – you’ll recall that virtually instantly.

The “Tweets per day” information is there because that’s what we’re really after. Let’s say you recognize someone, but their tweets per day is off the charts. And you can’t really recall engaging with them or being particularly moved by any of their tweets. Suddenly, you’re presented with a golden opportunity: Clear up a TON of attention span by dropping someone who is not particularly valuable to you.

The “Follows?” state is just to let you know whether you’re in a reciprocal relationship with them or a one-directional one. Odds are, if you’re in a reciprocal relationship with someone, you’ll be less inclined to drop them. They’ve accepted your noise as signal; and in turn you may prefer to retain them as a colleague or friend.

The layout of the page is intentional. Odds are, about 10-20% of the people you follow are not adding very much value. The more people you follow, the more likely that is the case. But it’s tough to decide that in isolation. Rather, the idea behind this layout is to provide you with a 1-in-6 opportunity to spot the weakest link. Think of it like a Twitter version of Jack Welsh’s Vitality Curve: 20% of the people you follow are A-players – you’re probably engaging them frequently; 70% of people are “B” players – you read them, you smile, but you don’t do much with them; and 10% of the people you follow are C-players – and not worthy of your limited attention span.

Of course, if you’re in a gambling mood, you can also play Russian Roulette – and 1 of the 6 people shown will be randomly chosen and un-followed.

So that’s it.

That’s the story behind TwitShowdown. If you’re getting too much noise, and not enough signal on Twitter, I welcome you to come play around with it. My hope is that by cutting out some of the noise in your Twitter stream, you’ll actually get more value out of it. If you’ve got any questions or thoughts, I welcome you to share your comments below. Thanks for reading!