Data Scientists for Insight

Today in a different Columbia course, we had the former head of Data Science from Air BnB join us to discuss some of the ways the company was using data to further their business. Interestingly, Air BnB has gone so far as to implement a "Data University" within the company to get all employees familiar with data and to a certain level of understanding. This type of fluency is probably one of the reasons for their success over time - both in their core product as well as their various marketing and recruiting endeavors for hosts and travelers alike. One comment she made that I thought was particularly insightful was around tension that sometimes comes up with data scientists and product managers when PM's ask for more tactical data and not true insights. I think this can be a problem in every company and particularly in marketing, where a marketing manager may ask for a certain data point or metric that they think tells the story they want to tell. However, instead, if the marketer (or PM, or anyone else for that matter) instead approached the data scientists with a more holistic scenario and set of questions that they want to answer and gather insights on, I bet the result will be tremendously more useful and beneficial to the company. So while we all want to brush up on our data use and KPI tracking, I think starting with the overarching question, or set of questions and then let the data scientist develop the right metric and experiments to prove/disprove will be more valuable. It also underscores the importance of tracking the right thing for your campaigns, which can be more difficult to identify than one may think.