Applying Fast and Slow Thinking to Data Analytics
3. Really consider the opinion of the most annoying person on your team.
Whether that person is Bob the Booer or Nick Not Possible, you need to listen carefully to their opinions.
Some of the best out of the box thinking is done by those who think differently than you do. And the negative thinkers are often the ones that can see flaws faster and help to spot issues that might arise when you launch models or algorithms.
Listen to them and make sure that you consider what they say.
This really serves two purposes:
1) It gets you a better result
2.) It makes them feel a part of the team. You just need to ensure that they don’t take the team down the proverbial rabbit hole where nothing happens.
4. Don’t leap to “causation” with correlated information.
Early in my career, I worked for a VP of Sales who was adamant that the analyses we did should not ever be construed as causation.
He was clear that we had correlated the rise in revenue to certain actions that had happened in marketing but was clearer still that it was not those marketing actions alone that caused the rise in revenue. When we are thinking fast, our tendencies are to believe the context with which we are most familiar.
So, if I am a marketing person, I can’t help but believe that the actions I take are the most important and therefore have “caused” increases in revenue. This is when thinking slow is very helpful.
5. Listen
The hard part of ensuring that your fast thinking consults your slow thinking lies in listening to those who understand the process of solving the problem and the data associated with it.
As analysts, we often are quick to provide information without realizing how the information we provide will be – or will not be — accepted.
Truly effective communication requires a sender and a receiver who both recognize the information being exchanged. And don’t forget that true communication starts with sender and receiver operating with the same language.
I know this sounds strange, but if you are in finance trying to communicate to sales or marketing leader, it’s always preferable to make sure that you have the same definition of data objects. For example, how is “discount” viewed in finance versus a sales or partner definition of the term?
What is an “account” to marketing versus an “account” to sales?
Takeaway:
As Kahneman explains, your mind is an incredible how it processes data into information. We, as mere humans, must strive to understand how we operate and to use our brain power in the most effective way for all. As we develop more applications that use artificial intelligence, let’s remember this concept of thinking fast and slow and incorporate into our models.
Related articles:
