Geospatial Visualization in Business: Bringing John Snow into the Boardroom

February 7, 2012

John Snow was a physician in London in the 19th Century and he is famous for having used maps to identify the source of the Broad Street cholera outbreak in 1854.  Dr. Snow’s work is often cited as the founding event in epidemiology.  For me, it represents a key event in applied geography, demonstrating the power of maps and their ability to illuminate patterns in data.

Here’s another map of the same data using modern cartography.

Just like the citizens of London who were trying to figure out what was causing the cholera outbreak, business analysts throughout the world are trying to diagnose pain points in their business.  Why are we having trouble keeping SKU#123 in stock in Enid, Oklahoma?  Why is our store in Paducah, Kentucky under-performing?  How can we reduce our transportation costs?  How can we avoid costly product returns?  Etc, etc.

Unlike John Snow, the vast majority of business analysts aren’t using maps to visualize their data.  Instead, they stare at row after row and column after column of data on spreadsheets in hopes that they will find a needle in a haystack.  Even if they do manage to find that needle, to communicate their idea may require getting colleagues to stare at the same rows and columns.  Sort of like looking at one of those stereograms where something pops out if you stare at it long enough – but for some people it never works.  For Seinfeld fans, remember the episode with Mr. Pitt (sadly, the actor who played Mr. Pitt, Ian Abercrombie, recently passed away) trying desperately to see the spaceship?

Imagine a group of military leaders standing around a table staring at a pile of spreadsheets.  Would that army win the battle?

So why do most businesses run on spreadsheets instead of maps?  Okay, I’ll grant you that for most reviews of financial data spreadsheets are the way to go.  I’m not saying do away with spreadsheets, I’m saying complement spreadsheets with maps (and other data-rich graphics).  This will do 2 very important things for your business:

1. Maps will help illuminate patterns in data that can’t be found with spreadsheets.
2. Maps allow almost everyone to contribute to the analysis and the discussion.

Let me give you an example from my own work.  In 2007 I was doing some consulting for a  national distribution company.  We were looking at the Atlanta market and trying to figure out how to optimize the configuration of branch locations to simultaneously maximize revenue while minimizing delivery times and transportation costs.  Most retailers and retail location research firms only look at maximizing revenue without thinking about minimizing distribution costs; but, that’s a subject for another blog post.  The group in the meeting room consisted of the CEO, VP Finance, Regional Manager, several Salespeople and several Branch Managers.  Anyway, we began by talking about financial performance for each branch location – the key concern was that, while revenue was strong, profit margins were too thin and needed to improve.  The VP Finance and the Regional Manager were running through PowerPoint slides that showed comparisons of revenue and profit by branch and making recommendations about which branches should be retained or closed based on historical financial performance.  Afterward, the Branch Managers and Salespeople, the “boots on the ground” in the Atlanta market, were asked to comment.  Their body language clearly indicated that they weren’t happy about the direction of the conversation but they didn’t have much to say.  In the meantime, the CEO was frustrated because he could sense the Atlanta team wasn’t enthusiastic nor forthcoming in sharing their thoughts on an important decision that was going to affect them more than anyone in the company.  He wanted their input.

After a short break, my colleague and I were asked to present our analysis.  We had prepared several maps for the presentation: historical and projected demand by ZIP code, locations of potential customers in the Atlanta market, and, of course, the branch locations currently operating.  In addition, we showed “sales penetration” by ZIP illustrating (roughly) the share of business that the company was achieving in each ZIP within the Atlanta market.  Finally, we showed a map summarizing delivery activity for the previous fiscal year.  The map showed the source branch locations and the destination ZIP codes.

Several mouths in the room dropped.  It turns out that each branch was delivering all over the Atlanta market and there was no cooperation to speak of between branches.  We also displayed a map showing the same data optimized for delivery efficiency and a reduced number of branches.

This map showed one way that the company could save nearly $3 million per year that would accrue directly to the bottom line.  In addition, delivery times would be reduced an average of 5 miles per delivery so customers would receive their shipments much faster (if you have ever driven in Atlanta traffic you’ll know that makes a big difference).

Now the discussion really got going.  It wasn’t long before the branch managers were on their feet pointing at the map on the wall.  The sales guys were up there too asking me to change slides so they could show the Regional Manager a pocket of demand that they’d been preaching about for years.  It was revealed that the new branch manager incentive program, which focused exclusively on revenue (rather than profit), was creating the unintended consequence of branches trying to keep as much delivery business as possible rather than share it with their sister branches.  Bad for the customers and bad for the bottom line.

You might be thinking, “yeah, but this analysis could have been done without maps and the results would have been the same”.  I disagree.  Again, the use of maps provided two advantages that helped facilitate a more productive discussion.

(1) The Baseline vs Optimized maps helped illustrate the point without having the exact number of shipments, dollars, etc on the spreadsheet distract from the overall message.  Everyone in the room immediately recognized the branches were delivering too far into each other’s territory.  A good map allows you to cut to the chase.

(2) Perhaps more importantly, everyone in the room felt comfortable participating in the discussion.  When the VP Finance and the Regional Manager were walking through the spreadsheet no one was willing to share their opinion because they didn’t want to look stupid for not understanding how to pronounce EBITDA.  Once a map was on the wall, the branch manager stood up right in front of the CEO and told him in no uncertain terms why his branch was losing money and how it could be fixed.

Pretty cool, huh?

Now, most geographers don’t produce maps that save a community from cholera (although health geographers and GIS-savvy epidemiologists do something very similar) but, like John Snow, they know that a geographical perspective may reveal solutions to a pain point.

What could a good map do for your business?