Supply Chain Expertise and Technology Blog by TMC, a division of C.H. Robinson

Benchmarking the Benchmarks


It is one of the most widely used freight network management tools, yet benchmarking is often misinterpreted and/or misapplied.

Understanding the application of benchmarking can help managers to avoid these pitfalls, and evaluate the performance of their supply chains more effectively.

In this post we take a look at some of the issues managers need to be aware of when assessing and using benchmarking tools. Next week we’ll delve more deeply into how these analyses can be misconstrued.

On a general level, be cautious about benchmarks that make sweeping observations about markets. Specifying the freight percentage of sales or what the cost per hundredweight of LTL shipments should be in an industry may be interesting, however, these values could pertain to average shippers – and that may or may not describe you.

In retailing, for example, a large bricks-and-mortar chain and a high-volume online player are both major shippers, but they differ greatly when it comes to freight mix and service goals. A leading big box retailer that ships primarily truckload has a freight percentage that bears little relationship to that of an online outlet with a lot of parcel shipments. In fact, one could be as much as ten times the other.

The point is that your metrics may vary hugely from the industry norm owing to the nature of your business. Or, it may seem that you are well positioned within the acceptable range for the industry, but the one-size-fits-all benchmark based on averages is a false measure of how you are actually performing.

It’s incumbent on the freight professional to choose and apply the benchmark that best fits the organization’s business model.

There is a useful analogy with car buying. When researching the price of a vehicle a buyer can consult various industry indices to obtain a market value for the target car. This value is based on a range of factors that are specific to geography, model, and make, as well as the features that the buyer requires. The purchaser could also use the average car price in the United States as a measure of the target vehicle’s worth, but this benchmark may be unrealistic because it reflects the price of every vehicle sold in the country.

Beware of unqualified averages. Scrutinize benchmarking consortia that are coy about the range of information they use to calculate average performance metrics. It may be that the number of companies sampled to support the findings is too few to be statistically viable. Ask about the standard deviation at the high and low ends of the range.

Also, be alert to the vagaries of survey-based benchmarks. The results tend to be biased because almost everyone rates themselves as above average when surveyed.

Check out the granularity of benchmarking data. Measures based on lane-by-lane or corridor-by-corridor, rather than network-wide, comparisons generally provide a more comprehensive and actionable picture of relative performance. Some model-based benchmarking services  offer this type of fine detail where others do not. Moreover, make sure the data base of information for the benchmark is big enough to ensure that the model is a realistic representation of the market.

More traditional corridor-based benchmarks offer the same granularity as model- based alternatives but can be flawed.  They may lack data in many corridors and only cover one or two shippers in others.  Or, to compensate for these flaws, the benchmark may include corridors that are too large to be meaningful.

Hopefully this brief tour of benchmarking pros and cons will help you decide which services best fit your business model and objectives.

If you would like to further discuss model-based benchmarking services, please feel free to contact me at

Next week we’ll consider some of the misunderstandings that muddy benchmark applications.