Using Transactional Data to Estimate Truckload Market Conditions in Near-Real-Time

Alex Scott
Read Time: 2 minutes apprx.
real-time analytics supply chain transportation

What is the status of the truckload transportation market?  Does demand outstrip supply?  Does supply outstrip demand?  Is the market relatively balanced?  These are hard questions to answer because, unlike commodities such as corn or beef, there is no centralized exchange for truckload transportation.

So if a shipper wonders why their contracted carriers are rejecting their freight, it’s hard to ascertain whether it is truly a “capacity issue” or whether their carriers are actually servicing other shippers’ freight (who may be paying more) instead of their own.  Or a carrier may wonder, should I allocate all of my capacity to a shipper with whom I negotiated a not-very-good rate?  Or is there likely freight out there for some of my assets that is more profitable?

In a recent project, I proposed a method to estimate truckload market conditions based on privately-observed spot bids from brokers and carriers.  The situation is thus: a large national shipper, when needing spot capacity (usually due to freight rejections), offers single loads to tens of brokers and carriers to bid upon via an auction mechanism.  When the shipper receives these bids, they observe the price that a carrier or broker is willing to accept to haul the freight.  Thousands of bids are received each week for various lanes around the country.  Some of these bids come from brokers, each of whom have relationships with hundreds or thousands of carriers.  Thus, if intelligently analyzed, these bids give us a signal of the status of the market.

The method to estimate market conditions requires multiple regression analysis using statistical analysis software (I use Stata).  It also requires a fair amount of operational data on each spot bid, such as the historical contract price, the lane, the pickup day-of-week and hour-of-day, among other variables.  But once these are all accounted for and I run an analysis, I then see a good estimate for the status of the market.  The advantage of this method over, say, the CASS index, is that CASS is run once a month and released about ten days after each month is over.  Also, it gives one number nationally, so it does not provide insight into regional conditions.  My method of estimation is near-real-time, to the point that (as of Monday) I know the conditions for last week.  Moreover, estimates are calculated regionally, providing deeper insight into the market.