Jack Levis, is known to many readers as “the UPS guy” in the Penn State Public Broadcasting production The Geospatial Revolution. He gave the opening keynote at this year's Location Intelligence Conference in Washington, DC. He basically backed up his now-famous quip about how UPS used to be a trucking company that used technology, but is now a technology company that happens to use trucks.
He reviewed what analytics are by highlighting their use in “data driven insight.” Organizations focus on first gathering their data together and putting them in order, gaining insight from them and making predictions with them. Said another way, they take data, turn theminto information, then to knowledge. Analytics come in three flavors:
descriptive - where we are
predictive - where we will be
prescriptive - what we should do
As use moves through these “levels”, more data, more expertise and more return can be expected. UPS captures more data both by using is DIAD mobile terminals (described as driver's "assistants" and four location data points captured for each stop. UPS also captures data about its vehicles and has learned that a vehicle that is an outlier in its performance is likely headed for a breakdown. Those vehicles are taken in for service before their planned maintenance cycle. Levis compared it to the Sesame Street game “one of these things is not like the others.”
Implementing analytics technology at UPS required updating every UPS procedure, including the details in the now 74 page drivers manual. Those changes have meant reducing 85 millions miles driven/year, 95% less time spent on training, 8M gallons less fuel used, 85K fewer metric tons of CO2 expelled.
The future looks bright. Even now customers can customize (for free) their delivery preferences (leave the package if I’m not home, leave with a neighbor, etc.) and even designate a delivery window within two hours. Levis and his colleagues want to go beyond prescriptive analytics to what he calls "clarovoyance," knowing what problems will pop up before they do.
Also in the future: a Nova episode with David Pogue that puts a spotlight on UPS’ use of analytics. Levis shared that Pogue did not stand a chance when competing with his routing optimization tools. In fact, he "fake fired" Pogue!
The Q&A was far-reaching but two questions, actually their answers, were startling. The first was about UPS use of real time data (weather, traffic, etc.). It currently does not use any. Why not? Adding it in along with the complexity of just its stops is too much. If real time is added, it will be real time data about if trucks leave a stop early or late. The other question was about UPS plans to sell its very detailed and accurate address data. While discussions have occurred and continue to occur with UPS mergers and acquisitions team, Levis hinted very few industries need the detail it does (he did note working with some emergency response teams). Further, he observed, trying to sell or license data people think should be free, doesn’t work to well.