While some concepts are appropriate for within-industry comparisons, such concepts are characteristically more concrete and objective (e.g., in the airline industry – number of lost luggage or number of on-time arrivals and departures, in the educational testing industry – number of points improvement in test scores, and finally, in an R&D organization – the number of patents achieved or the number of patent citations achieved). Industry-specific concepts are best benchmarked with non-competitors if possible or with competitors as a last resort.
Using competitor and industry reference groups are also problematic because data may not be available for all competitors or all industry members. In addition, many companies operate in multiple industries, making selection the correct industry reference group difficult.
Although scores on the MRI may indeed differ by industry to some degree, customer culture is not an industry-specific concept and therefore should be benchmarked across industries.
There is no benefit in limiting a company’s comparative group to its own industry or to a singular best-in-class company. Rather, concepts such as those measured by the MRI are best benchmarked with the group all of the companies for which MRI data are available, a broad cross-industry group.
It should not matter to which industry the “leader” on each dimension belongs – the high scores represent what is possible, what is best-in-class, and should be the long-term target for every company despite industry memberships.
An additional benefit of using a cross-industry reference group is that the benchmark is made in comparison to a much larger group of companies and scores are more likely to be accurate and stable as a result.
Conclusion
The MRI measures an important cross-industry concept. It is most appropriate to benchmark scores on this instrument in three ways:
- · Compare groups within a company to each
other.
- · Compare a company (or groups within a
company) over time.
- · Compare a company with all other
companies in the database.