2013-06-25

Co.Labs

How Aziz Ansari Tests Jokes With Analytics

The popular comedian is using the Web to collect data on audiences so he can test jokes on them. What’s to stop him from knowing exactly what his audience wants to hear, before he ever steps foot on stage?



This month, comedian, and TV star Aziz Ansari began inviting fans to “a series of intimate shows to work on new material for my next hours of stand-up comedy.” Called “Work In Progress,” the show will have tickets that cost just $10 and are awarded via lottery, ostensibly to reduce scalping and fan disappointment.

Comedians put on these kinds of test shows all the time, of course, often by dropping into comedy clubs unannounced. There’s something new about Ansari’s approach, however. To enter the lottery, fans must enter their name, age, gender, and relationship status, and Ansari implores fans not to lie. Why? Because he’s using the data to selectively test jokes on different types of audiences.

One thing I found very interesting is how differently these conversations are among different groups of people. Single people between the ages of 18 and 25 view these topics in a much different light than, say, single people over 30. Or married people over 40.

This gave me an idea. What if I could set up small shows to talk to very specific groups about these topics? What if I could do a show with half an audience of younger people and the other half is older married people? What if half the audience was single women over 30 and the other half was single men between 18 and 25?

By knowing the approximate demographic makeup of his audience, Ansari can decide to tell specific jokes each night that he thinks will work for a particular audience. He can then gauge their feedback and adjust his jokes accordingly to make sure audiences of all ages, genders, and marital statuses respond positively to his shows. It’s not unlike a developer using visitor segmentation to A/B test different calls to action, buttons, colors, and layouts on a website.

That got me thinking: What’s to stop Ansari from using those tools, too? Right now, the WIP shows are limited to New York and Montreal, which might be too specific an audience for a comedian with an international following.

Imagine if, instead of simply being a sign-up form for Ansari's WIP shows, the page were used to test jokes. He could promise fans an exclusive clip in exchange for entering their demographic data then serve different videos of him working with new material based on the info the visitor entered--one video for single millennials, another for married baby boomers, etc. After watching the video, the fan would be asked to rate the clip. Of course, reading ratings probably isn’t as valuable for a professional comic as taking the temperature of a room, so Ansari could even get really creepy and ask fans to let him record their reaction on audio or video. Then he could use machine-learning algorithms to analyze the recordings for laughter, pupil dilation, and even heart rate changes!

On second thought, maybe he should just stick to asking users to rate videos given what typically happens when you let people submit anonymous videos. In any case, by using the Web to collect data on his test audiences, Ansari is, however slowly and basically at first, starting to take comedy into the realm of data science by trying to understand who his audience is and what different segments of them like. If he succeeds, it’s only a matter of time before other acts and creative professions follow suit. Let’s just hope that the data doesn’t wash away the creativity entirely. That would be ironic, but not particularly funny.

[Image: Flickr user Andreas Bradt]


Article Tags: comedy does data





Add New Comment

5 Comments

  • neugenes

    Great, now I can dislike Aziz Ansari even more! He's always given off that "everyone please love me" vibe. Now he's just taking things to the next level.

  • The World Is Depressing

    Well, I'll officially never be watching or supporting any of his "comedy". Thanks for the heads-up.

  • Ankur Oberoi

    with WebRTC in the browser, the machine learning parts you mention could be possible. the company I work for, TokBox, will be working on enhancing our existing API, OpenTok, so that developers can record videos in the cloud. this would be the perfect building block for someone to come and plug in some data analysis algorithms and presto!

  • Adam

    I went to one of these shows about a week ago, and you could tell that the younger, single people were seated close to the stage to more easily facilitate the interactive part of the show.