top of page
  • Ekow Duker

Laughing Out Loud




As the year draws to an end and we enter the festive season, many of us will be looking forward to spending time with loved ones in an atmosphere of happiness and laughter. Making people laugh however, is no joke. Here are four lessons from the world of stand-up comedy that bear surprising similarity to the art of data science.


1) Experiment all the time

[Stand-up comedy] is a true “learn-by-doing” art form and you won’t learn what works (and what doesn’t) until you’ve gotten on stage in front of an audience. The more chances you have to perform, the more you’ll be able to learn.” 

Patrick Bromley in 10 Tips for Beginner Comedians


And so it is with data science. While data scientists will spend hours crafting really smart algorithms, the value of their work is only truly known when their models and data pipelines are tested in the real world. There is simply no substitute for that.

This notion of experimentation can be difficult for traditional organisations to make peace with. Dour executives, accustomed to business cases with an iron-clad rate of return, may be missing out on the benefits of agile discovery that are necessary to keep up with today’s customers and their rapidly changing behaviour.


2) Choose your pilot projects carefully

Even very successful comedians such as Jerry Seinfeld, cannot come up with an hour of material without testing it on smaller crowds. They will often go and perform short acts of new material in small comedy clubs, many times without any advertisement for their appearance as they build their way into a full act.

Gil Greengross Ph.D. writing in Psychology Today about the The Fascinating Life of Comedians


Andrew Ng, the Founder & CEO of Landing AI and former Chief Scientist at Baidu, echoes a similar sentiment. In his words, “It is more important for your first few AI projects to succeed rather than be the most valuable AI projects.” This carefully, carefully, approach is particularly important for organisations new to data science for whom a failed first set of projects, can quickly dampen enthusiasm and support.


3) Really get to know your customers

“There's a distinct audience for every specialized group. They are categorized by hundreds of special interests: color, religion, education, financial and social standing, acumen, geography, politics, fame, and sex. The same material that works for a college audience will not work for a group of lawyers, doctors, or bankers.”

Comedy Writing Secrets by Mel Helitzer and Mark Shatz


Data scientists do something very similar when they deploy machine learning techniques to coral customers into homogenous groups or look-alikes in order to drive more meaningful customer interactions.


4) Think hard about the end-to-end customer experience

“It’s a musical flow; a comedic symphony of relatable experiences and observations bonded together through transitional phrases.”

Brandy Thomas, Comedian, New York, in How to write stand up comedy material


Ultimately, a data science program in a customer facing industry has as its goal, the creation of a joined-up set of meaningful experiences that collectively, nudge customers down a path of increasing loyalty and engagement. When done effectively, the up-side can be significant. And that’s no joke.

(Photo by rawpixel on Unsplash)
4 views0 comments

Recent Posts

See All
bottom of page