Using Big Data and Analytics to Manage Risk • 217
third- party tool for spend analysis and procurement performance analysis.
ese data were cross- referenced against an external database with thou-
sands of price indexes. is approach of combining internal and external
data and benchmark indexes allowed for fact- based discussions and deci-
sion making for continuous improvement in its cash ow management.
is approach conrmed that analytics works best when integrated
with external information. e utility company concluded that integrat-
ing internal and external data through the use of big data analytics is
an enabler for managing supply risk, supplier risk, regulatory risk, com-
petitive risk, and intellectual property risk. Managing these risks should
include analytical approaches such as scenario planning, Monte Carlo/
probabilistic modeling to quantify the probability of occurrence and
impact, segmenting and visualizing risk using heat maps, and predictive
analytics to manage risk.
Schneider
Schneider National, a $3 billion transportation and logistics company, has
developed a computer model that mimics human decision making, help-
ing the company to assign trucks and drivers in the most cost- eective
way possible. At any given time, Schneider has 10,000 trucks on the road
with over 30,000 trailers waiting to be picked up or delivered. Drivers work
alone or in pairs, and Schneider must get them back home by a certain
date and time. Drivers also need to conform to the government’s hours-
of- service regulations regarding rest periods and breaks.
With the help from several Princeton University researchers, Schneider
developed a simulator utilizing dynamic programming, which takes into
account the presence of uncertainty. e simulator, which took two years
to develop, runs forward in time for three weeks to approximate the value
of having a truck and driver at a certain location at a certain time. e out-
put from the report is a called a rst- pass cost estimate. e tool then runs
backward in time, something called postanalysis, to reconcile the past
results with those that were determined in the future estimate. e simu-
lator then runs forward again for three weeks and then backward as it
seeks to improve the total cost estimates. is forward– backward process
encompasses hundreds of thousands of iterations.
Schneider estimates its big data tool has saved the company tens of
millions of dollars as well as increased revenue by justifying price hikes
to customers with specic service- level constraints. e simulator also