Rage Against the Retweet Machine

Posted on April 9, 2013 by Kalle

The RetweetGate

Kudos for The New York Times for breaking the story on fake Twitter followers becoming multimillion-dollar business! We have noticed the same phenomenon with @TheNextWeb’s exceptionally high RT volumes using our twheel™ Twitter app, and have recently been collecting data on @TheNextWeb and it’s peers in the world of technology blogs to investigate how retweeting behavior differs between the sources.

White bars indicate the amount of retweets a tweet has received

twheel™ uses retweets for visualizing the traction each tweet has received from others. You can read more about twheel™’s design principles here.

We performed a set of studies by creating a simple twitterbot with Google Apps Script that pushed the retweet counts to a Google spreadsheet.

We ran the first study on March 20th, 2013, where we tracked 10 tweets from Wired, The Verge, Gizmodo, CNET, The Next Web, GigaOm and Engadget. The retweet accumulation samples were checked every 15 minutes for a 24 hour period.

To confirm our findings, we made another study where we compared 13-25 tweets from The Next Web and Techcrunch (April 3rd) and Pete Cashmore and BGR (April 8th). In both of these studies the retweet accumulation samples were checked every 5 minutes for 24 hours.

It doesn’t take a genius to spot the unnatural behavior of The Next Web’s retweet accumulation graph. While The Next Web published a lengthy take on why fake followers aren’t worth the risk of losing credibility, they clearly aren’t following their own advice.

The averages are calculated from the Tech Blog Retweet Face-Off (Round two) data.

What is the benefit of using bots?

1. Reach a wider audience

Fake follower bots help you get your message to reach a wider audience. In the New York Times’ article The Next Web’s Zee Kane talks about the whole retweet thing being the serendipitous result of experimenting with their own TNW Labs product spread.us, which claims to provide a service that lets you post tweets on behalf of your colleagues, friends, and fans. The aim of the service is to “Create massive social reach together”. This sounds like a social media marketers dream, but how do Cilia Poon (@health_c_f_shop), Kristan Elverson (@pops260) or Bashir Ako (@LadanDaura) become the colleague/friend/fan of @TheNextWeb? Are they true fans, bots, or even hacked accounts?

Instead of reaching just their 872 000 followers, with the help of their bot powered fake colleagues, friends and fans The Next Web can reach multiple times larger audience than they have earned.

2. Grow your social ego and influence

The amount of followers and the amount of discussion you can generate in Twitter are the key ingredients how Twitter itself and services like Klout.com define the sources importance and influence.

In Seth Stevenson’s excellent article he talks about his experience of buying 27 000 fake followers; ”I noticed that after I’d bought my zombie followers, the rate at which new, nonzombie people followed me seemed to rapidly accelerate.”.

Easy money. Fake it to make it.

Manipulating these metrics can also have a very potent money making effect. According to researchers quoted in the New York Times article the value of the fake followers market is somewhere between 40 and 360 million USD. Best fake follower providers brag of making upwards of 1 million USD a week. Is selling fake followers bigger business than Twitter itself?

Why doesn’t Twitter just kill the fake followers?

So why hasn’t Twitter taken action, even though spotting this type of unnatural retweet accumulation should be pretty simple? Perhaps, because over 80 million of Twitter users (40%) never tweet anything, so it might make business sense to have the 20 million fake follower bots make Twitterverse look active and social.

Also, it appears that Twitter has no way of recognizing bots from real humans. According to Twitter spokesman Jim Prosser “Forty percent of our user base only consumes content. What looks like a fake account to one individual could actually be someone who is on Twitter purely to follow people — like my mom, who follows me and my brother, doesn’t have a profile bio and has never actually Tweeted herself.”

Why are fake followers/bots so hard to identify?

Because, at the moment, social interaction is measured by tracking actions (retweets, tweets, replies, favorites and link clicks). All these tasks can be easily performed by a simple computer program. With a simple script and a little help from cheap indian labour, or perhaps even the elderly . There is no way to check if the actions are being performed by a real human with real motivations, or generated by a simple bot.

Is Social Media doomed to be overtaken by bots?

Creating fake followers is so easy that “a kid could bypass Twitter’s defenses”. At the moment, social influence is for sale. Buying followers and retweets is easy, cheap, and the risk of getting caught is minimal. Are we humans doomed to become a minority in the social media like Twitter? Is there anything we can do to stop the bots?

Yes, there is.

To identify a bot from a human, we need to start tracking attention instead of actions. This is what we are doing with our upcoming Twitter news discovery app twheel™ 2.0.

Reading behavior of a single twheel™ user. twheel™ tracks the time spent reading each tweet, creating a direct measure of user attention.

It’s time to fight back. The battle will start soon. Stay tuned to learn more @twheelapp @kallemaatta.

twheel™ -Against the Bots!

The Future is here!!

Posted on September 14, 2012 by Kalle

Dear Jolie O’Dell

Posted on August 22, 2012 by Kalle

We will keep you posted.

The design philosophy behind twheel™ – It’s time for a UI revolution!

Posted on August 15, 2012 by kride

Since twheel™ caught a wave of interest from the start, being referred to as “The Best-Looking Twitter App We’ve Seen So Far”, and “fun, gorgeous, and… absolutely stunning”, we thought you might be interested in learning a few things about the design, philosophy and thinking behind twheel. Here are top three golden rules we live by:

Rule #1: All tweets are born unequal
In communication, contextual information – who says and how the others react – can typically be just as important as what is actually being said. So instead of trying to ﬁnd interesting tweets based on the content, with twheel you can get the big picture of your timeline.

With traditional list-based Twitter apps you have to practically read through everything to pull the signals out of the noise. Depending on the number of sources you follow, the resulting signal-to-noise ratio can be frustratingly low. By visualizing the metadata related to each tweet (like RTs or replies), twheel makes it easier for your brain to find and consume the relevant tweets.

Rule #2: Human cognition and physiology matter
All of the current social media clients present information as endless lists of textual information items. Reading through lists requires the use of sharp central vision, which requires fixating each word individually. While our brains construct an all-around sharp perception of our environment, it’s just an illusion.

No, you can’t read multiple tweets simultaneously. Due to physical limitations, such as the number of photoreceptors on the retina, the eye can actually transmit “sharp” data only from the very central vision.

However, our brain is wired up in a way that makes us good at spotting shapes, patterns, forms and colors, and doing this even with peripheral vision. A recent study demonstrated that people can spot shape deviations in multidimensional polar diagrams in under 100 milliseconds and we can even specify the location of the deviations within the form with surprising accuracy.

Studies like these brought about the use of shapes and deviations, colors and contrasts in the design of twheel. This, combined with the need to condense information in the small screen space of mobile devices, made us forgo existing conventions of data display and create something that is not only new but is more intuitive and easier for our brains to consume. Since our team boasts not one but two cognitive researchers, we can safely back up this claim with a simple: “Trust me, I’m a Scientist”!

Rule #3: Touch is different
Selecting targets on small mobile displays is another fundamental bottleneck of interaction due to limits in pointing accuracy. We’ve all probably noticed at some point that our fingers were a bit too big for our mobile device’s keyboards! Thanks to the evolution of touch screens, we can now ditch the old ideas of human-computer-interaction of making the screen into a grid of clickable objects (which is a desktop convention from mouse-controlled WIMP interfaces and start developing more context-aware and brain-ﬁendly solutions, like twheel.

In a typical mobile use scenario, the most optimal way of holding and operating the device is with a single hand. However, it’s often hard to navigate the touch screen with one hand while keeping a firm grip on the device. Evolution has equipped us with just one opposable thumb, and reaching targets across the whole screen can be cumbersome.

With the twheel’s rotating wheel, we can bring all the targets within reach, and place a practically inﬁnite amount of targets under your thumb!

This is twheel™

Posted on August 7, 2012 by Kalle

The reason we decided to (r)evolve Twitter with twheel is to provide better tools for discovery.

The volume of information we receive constantly is overwhelming. We need to be able to find the relevant items more effectively. As Twitter CEO Costolo put it: “We’ve got to ﬁgure out how to capture the volume at the same time as separating the signal from the noise.” [ ].

At the moment, the giants of social media are trying to solve the problem of information overload by reducing the amount of information through filtering: Twitter with their tailored trends, Facebook with highlighted stories, and Google+ with personalization. Filtering, however, will eventually put you in a filter bubble (a term coined by Eli Pariser in his TED talk), where an algorithm decides what you are able, or even allowed, to see.

We decided to take the other route, and follow the words of Edward Tufte: ”information overload is not caused by the amount of information, but failures in design.”. That is, to solve the problem of information overload, we don’t need to reduce the amount of information through filtering, we need to change the design!

We selected Twitter as the first social media ecosystem for twheel, because we love Twitter and it’s open nature. We’re striving to make twheel the most beautiful Twitter client out there. And not just to please your eyes, but to fondle your visual cortex as well. It might take a while for you to get used to twheel, but once you do, you’ll love it. We promise!

twheel™ is now available in the AppStore.

Cheers and enjoy,
Team Fluid

Coming next: The design philosophy behind twheel.

twheel is a registered trademark of Fluid Interaction Oy.

Bots Checked-In at Social Media and Ousted Humans

Posted on July 4, 2012 by kride

A [ro]bot may not injure a human being or, through inaction, allow a human being to come to harm.

– Isaac Asimov’s 1st Law of [Ro]bots

Apple, the force is strong with you. Thanks to you, applications are becoming more and more social. First came the close integration between iOS and Twitter, which tripled the number of iOS users on Twitter (10 billion tweets and 47% of all photos on Twitter come from iOS 5). Now the same thing is happening with Facebook. The social network is already spewing out billions of messages a day. Will iOS integration triple these as well?

A recent study shows, however, that most followers of major brands are just bots. Where does this leave us humans? Certainly not any closer to the joyful accidental discovery of gems, or serendipity, that Google and Twitter are selling us. Rather, this leaves us more lost and insecure under the ever-increasing information overload.

All social media channels are boasting their success based on the amount of messages they deliver and the new connections they build. But how valuable is a connection between an app and a bot? We believe that in this case quality definitely beats quantity.

We are fed up with getting all the unnecessary automated messages from some app or game a friend is playing. Imagine what would happen when home appliances finally get their promised IPv6 addresses and their own voice in the future. Won’t you be thrilled to get messages like: “Bill just flushed the toilet”?

Google, Facebook and now even Twitter are trying to solve this problem with their huge server farms, personalizing information for us based on our previous behavior and connections and bringing the content that they think we like to the top of the list.

But by doing this they are just adding a new level of bots and algorithms between us humans and the information we desire. In the end, how much does a programmed algorithm differ from Internet censorship imposed by an oppressive government?

We don’t like the idea of some middle man telling us whether we should be interested in something or not, or even worse, censoring part of the messages we receive. When bots are in charge of selecting the information we are allowed to see, Google’s own dogma “Don’t be evil” rings hollow.

We want our social media to be our own. We want them to consist of people we decided to share them with and be free of outside filtering.

This is why we are developing Twheel [http://www.twheel.com]. With Twheel we are using the most powerful super-computer ever created – your own brain – to make your Twitter feed truly personal and help you find important signals easier.

Say “No” to bots and “Yes” to human-powered solutions!

Fluid Interaction blog

Signals from the noise and other stories about brain-friendly information design