paint-brush
How I got myself banned on Twitterby@jsemrau
167 reads

How I got myself banned on Twitter

by JanFebruary 5th, 2017
Read on Terminal Reader
Read this story w/o Javascript
tldt arrow

Too Long; Didn't Read

It was all our fault. Mea culpa. We vastly underestimated how bored people are on Twitter and how many people will post on Twitter about being bored.

Company Mentioned

Mention Thumbnail
featured image - How I got myself banned on Twitter
Jan HackerNoon profile picture

Why is everyone so bored?

It was all our fault. Mea culpa. We vastly underestimated how bored people are on Twitter and how many people will post on Twitter about being bored.

It all started a couple of days ago when I learned about a powerful new mechanism to potentially increase customer awareness called “social listening” and thought I’d give it a shot.

Why? One of the main problems startups have is to get noticed in the noise of online chatter. Tenqyu is no exception here. It is quite tough to compete for attention with the frequent Trumpisms, the Superbowl (Go Raiders! ), Memes, Gifwars, a lot of online communication of 2017 that is shaped by bots talking to each other .

So in order to get seen, I figured, it would be sensible to automate one headcount in the social media team and run a bot on my own.

Implementing the bot from python was not hard and is actually quite straightforward. After quickly outlining the process, the team in the innovation lab connected the different parts and we rolled with it. #weekendhackathon

As the good citizens we wanted to be, we thought we were following the Twitter guidelines by offering a service similar to this:

Sunday afternoon we had the system ready to go live for a first test run. As drop!in is focused around geo-spatial aggregation we needed to identify the location of the tweeters and then create a custom event notification for this specific user. Similar to this one.

When we were looking at the candidates during the model building we noticed that some of the tweeters had not put in a proper location. So, in good engineering practice, we defined that in this case the system should reply with a default link to the app. (Hint : Big Mistake)

From our calculations and the sample data before we estimated that the default case should occur only in roughly 5% of cases when we can not establish the location of the user.

Then we set the system live.

Within a few minutes we had informed several users about the app but not a single event. Bummer. Apparently our fail-safe option has become the new normal. Sheeeeet! Why are so many people tweeting about being bored on Twitter ! WTF!

Overall roughly 40+ Tweets went out when the app was greeted with a lovely error 261 “ Application cannot perform write actions”.

Twitter shot down the app restricting write access within no time. Well deserved, I suppose.

But the solution proves immensely powerful already and thus was a big success. For us this is clearly the way forward for the application and once the app gets reinstated we will remove the default option as, it is agreed, that it does not add value and surely does not add a “whoa ! This is amazing effect” rather than a “WTF SPAM MONSTER!! BRRRGH!!!” . And surely this is something neither Twitter nor we want.

Maybe we move from “I am bored” to the search keyword “event suggestions”.

But I keep this for another post what will happen here.

Which keyword would you be using?

Thank you for coming here. Please let me know if you liked the write-up :