Starting in 2000, I invested more than 7 years of my career at a company that sold mobile fleet management software to the mid-tier services market. That was the time when modern logistics was in its infancy. I developed on products that would use software wizardry to find low-cost combinations of drivers and visits and spit out routes and schedules. The big boys, like UPS, were just figuring out that avoiding left turns saves a lot of money. It seemed like my company was ahead of the curve.
At the time, our special sauce was a combination of a geocoding engine, GIS database and a program that calculated the cost-reduced schedules using a mathematical model of the world. We even used an early version of MapQuest’s mapping API. Nearly all of the under-the-covers components and techniques we developed or used were powerful, yet difficult to explain to the layman. Limitations in computing power made selling to large MSAs a challenge. The implications that our customers’ businesses had to change to adhere to the model was even harder.
Fast forward to today. Google has made geocoding, multi-point routing and getting turn-by-turn directions an everyday activity. All of our gadgets have a GPS on them and there are dozens of AVL packages that tap into them. All of this commoditization has turned field logistics into a big data problem.
With the off-the-shelf software libraries (many of which are open source) that do a “good enough” job addressing the geocoding, mapping, routing and calculating, companies playing in this space must distill their special sauce down to one thing: their economic model. Technically speaking, this is the functions and equations that determine the cost of driver A driving to points X, Y and Z go far beyond fuel cost per mile. Factors like “being late,” “being liked” and “having the right parts on your truck” included in the model make it more realistic. In order to compete, field logistics software must effectively use all of the data it can to maintain a pragmatic, real-world view of the logistics business, so that decision-makers can efficiently make trade-offs.
However, it’s still difficult to sell a software package or service on “the best” economic model if it’s not consumable or affordable. It’s like selling a dream that you don’t quite understand. One lesson we’re still (hopefully) retaining from the dot-com bubble burst is that you can’t effectively sell a product based on a simulation alone. In today’s world of try-before-you-buy software consumption cycles, it’s critical that prospects be able to easily kick the tires on the model before plunking down big bucks on the bun and the patty.