Amazon Prime Air — An Analytics Metaphor

I’m assuming most of us saw Jeff Bezos’ announcement on Sunday about Amazon Prime Air — an ambitious plan for Amazon to deliver packages in 30 minutes via quad copter.  Sure, it may have been a PR stunt and it certainly got some good natured ribbing from the internet, but it’s definitely one of those things that makes you think about business models.  Think about going from next-day delivery to same-day delivery to 30-minute delivery…now all we need are the damn Heisenberg compensators and we’re all set!

Amazon Prime Air Quadcopter

From an analytics standpoint, it made me think about a concept that I repeat often — get the data in the hands of someone who can do something about it.  In Amazon’s case, the goal of 30-minute delivery is to improve customer satisfaction.  The quicker the customer gets the product, the happier they are.  But it also accomplishes a different goal.  It reduces the friction (time, cost, effort) for the consumer, thus making the decision to purchase from Amazon an easier one (as my colleague Andy Allen says, they’re building the infrastructure to put a warehouse stocked with 100 million items right in everyone’s pocket).

Analytics (or data or reporting or business intelligence) works the same way.  If a worker needs to wait for a report to run or spend time converting a report into something that’s more usable, that adds friction to the process.  Adding friction will usually end up in reducing the overall usage or utility of the data.  That’s a shame…I hate it when good data go to waste.

A good analytics project should not only focus on the product (the report…or the drone’s payload), but it should also focus on how to get the useful information into the hands of it’s intended audience in the most expedient fashion.  Here’s an example.  In higher education, there are many folks who can generate a student risk score.  “What’s the probability that this student will not complete this course”.  That’s wonderful information.  If used correctly, the student might get the help he or she needs before dropping off of the radar altogether.  So what should the institution do with these data?  A report listing the students and their risk scores is nice, but it doesn’t do much good if it sits on the desk of an administrator.  One path is to push a report out to the faculty member.  Another is to notify the student directly.  Yet a third is to populate a CRM or a case management system directly.  The model says that these 48 students in the Psychology program are at risk?  Well, let’s create 48 tasks in the Student Services case management system ASAP so that advisors can start contacting students.

Thanks for the inspiring thought, Amazon.  Now where’s my flying car?!?!