Taking the Data Bus

As some of you may know, because I keep telling you, when I’m not writing witty posts for a limited reading public, I’m a dyed-in-the-wool data geek.  I’ve just come from a data conference in Cambridge, Mass and what a fun time it was!  Data everywhere you looked.  Even better, stories about what the data is telling us.  Then, fruity drinks and chats with fellow data geeks in the evening.  Pure heaven.

Again, not that kind of Data.

Data can tell you things that, apparently, people can’t.  For example.  After collecting lots of data about patients and doctors, one large pharmaceutical company came to a surprising conclusion.  They were collecting data about how well people comply with the doctor’s orders.  Here’s how it works:  your doc prescribes something for you.  Blood thinner, allergy pills, Viagra, whatever.  Do you actually fill that prescription?  Do you take it?  Or do you hide it under your tongue and spit it out later, like a cat with a pill buried in the tuna fish?  The first thought was, let’s see which meds people won’t take.  So, that counts Viagra right out from the start.  Then, the data scientists started looking at other “correlations”. 

I’m part of study about med compliance!

A correlation, for those who don’t know or don’t want to know, are things that happen together.  The sun rises and the rooster crows.  Those events are correlated.  Sometimes, the rooster crows and then the sun rises.  Still correlated, though you can’t say for sure which causes which. 

Anyway.  The “surprising” correlation was that if the patent is male and the doctor is female, the meds are taken way less often than any other combination of doctor/patent gender. “This was surprising!” said the older Midwestern white guy giving the presentation.  “Who could have guessed this?”  My left hand went up immediately, even as my right hand cradled my sadly shaking head.  How little we have learned, my friend.  How little we have learned.

I’m astounded by your correlation, sir! (thanks, Shutterstock)

I hope this brief incident may have clued Mr. Pharmacy into a basic fact of life, revealed through the miracle of data analysis:  my experience may be very, very different from your experience.  Mr. Pharma seemed to be a smart, honest, and good-hearted captain of industry from a medium to large city in the Midwest.  However, he had clearly never once considered that there are differences between how I go through life and how he does.  He had a brief flash of insight when, seeing my hand shoot to the sky and hearing the widespread female laughter surrounding him, he said “well, other than every woman in this room.”  Try “on the planet”, buster.  It’s there – in the data.  Learn from the data.

Thanks, paid promotional spokesperson!

So that’s your lesson about “correlation”.  Now, let’s talk about data and business planning.  Most of you think this is the kind of dry, financial information best held in a spreadsheet.  You think of guys in ties using laser pointers to show pie charts (mmmm, pie!).  But wait, there’s more!  Sometimes, your business plan isn’t taking reality into account.  Your data is missing something.  That can be a problem.  Remember that guy who got dragged off United Airlines?   Maybe the business plan didn’t have a data category for “dragging paying passengers off the plane”.    Management missed the cue, and it took some widespread social mocking to bring them up to speed.

Mmm, pie charts! Shame this one adds up to 271% total. Yeah, I know – top 3 pies. Should be 300%.

So now, we come to the point.  Does your business plan allow fresh data points to be added in?  Is it flexible enough that when a so-called “black swan” event comes along, you can respond and change?  I’ve always had a problem with the “black swan” analogy.  If I see a white swan, I’m not an idiot.  I can image black swans, pink swans, even glow-in-the-dark florescent green swans.  Plastic swans.  Pictures of swans.  That takes only the bare minimum of imagination.  What I can’t imagine, after looking at a white swan, is a platypus.  When a platypus comes along, do I throw it out of my data because it’s too weird?  Or, do I adapt?

Swans, flamingos — not really that different.

Here’s a real-life example.  After a lovely seafood meal in Cambridge (I liked the Summer Shack or try these places), I was ready to board the GoBuses bus back to Manhattan.  It’s cheap, convenient, and ready to go, as the name implies.   But.  When I boarded the bus ready for my return to my native land, data charts dancing in my head, my nose alerted me to something a bit….off. Not to mince words, it was a bag of vomit in the middle of the aisle.  I de-bused, and told the driver.

Not a swan.

His response, “well what do you want ME to do about it?  I’m the driver!” might have driven some people to a nasty reply.  Not me.  Fresh from an objective, data-driven view of the world, my response was “Well, I’m the passenger, so I won’t be cleaning it up either.  However, I feel strongly it’s a task that should be done, and soon.”  The customer service hotline informed me that the problem could be addressed once we got to Newton, 30 minutes away.  You don’t need Power BI to tell you that 30 minutes with a bag of vomit is better than 6 hours.  I positioned myself in the first seat, just behind the driver, to take advantage of airflow.  Because I’m a yenta, I warned everyone else as they boarded.  And, because I’m a yenta, when we pulled into Newton, the driver told me on the lowdown, “You’d better tell that guy in the green vest.  He’s in charge.”  “But I already called the customer service hotline!”  His snort of derision told me all I needed to know.

Vomit speaks many languages.

The guy in the green vest was the station master, and he was not happy with my breaking news.  Although he didn’t actually say “well what do you want ME to do about it?  I’m the station master!”  He did scowl strongly, and comment “I bet YOU’RE not going to lift a finger!” to the driver, who shrugged and said “it’s not my shift anymore,” then ambled off to find a sandwich (“who could eat lunch with that smell?” he asked philosophically as he walked away).  The fresh driver was standing at the ready, but it clearly wasn’t his problem either.  The station master found some poor overworked counter flunky to back him up with gloves, paper towels, and spray cleaners, and before you could say “why is this bus already 30 minutes behind schedule?” the job was done.

My thought experiment might get me to take the train instead.

And here’s where a good, flexible data plan comes in.  GoBuses has a great business plan.  The price is right – under $20 NY to Boston on a weekday.  The buses are in good working order.  They don’t have to pay rent in Port Authority because they stop on side streets west of Penn Station.  The website and ticketing is a breeze.  Search engines?  Tweaked.  GPS for drivers?  Industrial strength.  Equipment check?  Power plugs, wi-fi, bottle holders, and footrests.  Whose job is it to clean up the vomit?  Crickets.  It’s just not in the data that goes back to the head office.  No one ever thought of it.  The airlines – they thought of it!  Little bags everywhere you look, and discreet, thoughtful, be-gloved flight attendants whisking away the horror, the horror.  So, GoBuses – figure out whose job it is.  Maybe it could be as simple as stipulating to the driver “yes, it IS your job to clean the vomit, and you get fifty bucks combat pay for every incident.”  Include an emergency cleaning bag along with the emergency first aid kit on each bus.  Do something.

It’s not his job, either

Your customers will thank you.

Leave a Reply

Your email address will not be published. Required fields are marked *