Beware of biased benchmarking: 3 ways to grade a transportation index

Benchmarking different aspects of operations is a well established practice across most businesses. Understanding how your company compares to similar organizations in terms of efficiency, delivered level of service, cost or other dimensions allows you to identify trouble spots and focus your managerial attention.  

Benchmarks and indices are related in that a benchmark is simply a comparison to a commonly accepted standard. Benchmarks and indices, especially for costs of production, exist in many different markets, including pulp and paper, rubber and plastic, and most other commodities. Transportation is no different. The Department of Energy’s Diesel Fuel index is published each week and is the basis of most firms' fuel surcharge programs. Having an established, recognized and trustworthy index makes sharing fuel costs between shipper and carrier easier to manage.

Get the clearest view of the freight markets with benchmarks and analytics from the Freight Market Intelligence Consortium.

 

The usefulness of a benchmark, or an index, is strongly correlated to the quality and source of the data used to generate it. If the data is skewed or biased, then the index will not be helpful to the user; in fact it could lead to bad decision-making. This is especially true for benchmarking truckload transportation rates and level of service since it is such a large market (~$300B) with essentially an unlimited number of unique point to point moves that each could be priced differently. Whenever considering the use of a benchmark or an index, ask these three questions:

  • How are they calculating the benchmarks?
  • Where is their data coming from?
  • What are their goals and objectives?  

Having a valid transportation benchmark allows a shipper to understand how they compare to the general market, determine where they might be overpaying or being underserved, and help establish a transportation budget going forward.  

All transportation benchmarks are not the same, however. To illustrate this, take the FMIC as a benchmark. It is very transparent in its benchmarking method, data and goals. 

Methods

The methodology used to develop the FMIC benchmarks were developed in 2004 and have been run continuously since. The truckload benchmarks are determined by breaking down truckload rates into the key cost drivers (distance, origin, destination, mode type and other service factors) and using regression to estimate a market rate for every possible point to point lane.  This methodology allows for not only benchmarking low volume lanes, but also estimate the impact of different policies and practices on rates.

Data

Currently, the data used to develop the FMIC benchmarks come from over 30 million actual shipment transactions. This represents an annual spend in excess of 28 billion dollars from more than 130 shippers representing a wide range of sizes and industries. It is a truly representative sample of the North American truckload trucking market. 

Goals

FMIC and DAT exist to create greater clarity and visibility in the transportation market as a neutral provider of data and insight. DAT and FMIC do not broker loads or sell transportation management or consulting services. Our goal is to be an independent benchmarking authority -- a reliable source of truth for all market participants.  

How benchmarks can fail

Benchmarks that are weak or inconsistent in any one of these three dimensions (method, data, goals) lead to a bias that can inadvertently misdirect decision making. If a benchmark relies heavily on a single segment of the $300 billion market, it could lead to incorrect signals. For example, if a benchmark had relied only on data from the industrial sector over the last six months, it would have indicated a precipitous drop in shipment volumes and prices. Conversely, concentrating on the retail and CPG sector would lead to the conclusion that the market is over surging. Only by including a wide range of shipper sizes, industries and commodities can a true picture of the entire market emerge.  

Having mixed goals can also bias a benchmark. For example, if a firm provides procurement and management services to the same companies that it is using to calculate a benchmark, then it is biased. In other words, if you are measuring your service provider on how well they are managing your overall transportation cost but that same service provider is giving you the benchmarks to measure their own performance, there is going to be bias. These internally facing benchmarks do not capture the diversity of the true market and reinforce that company’s own buying habits and patterns. Beware of software providers who throw in ‘benchmarking’ for free as part of their service.

To avoid bias, the company providing the benchmark has to able to list the three questions listed above: How is the benchmark calculated, where does the data come from, and what are the company goals?

If they cannot answer these questions to your satisfaction, then perhaps you should look for a different benchmarking partner.  

Learn more about our FMIC and Pulse benchmarking services, or message us at Enable JavaScript to view protected content.


DAT LOAD BOARDS
Fastest way to
find trucks


Chris Caplice

Dr. Chris Caplice is the Chief Scientist at DAT iQ and FMIC, as well as the founder and co-director of the MIT Freight Lab. Chris also hosts the Freightvine podcast