More On

# Morrison on Metrics: Absolute Clarity About Normalized Metrics

An absolute number, such as 10 lawyers, differs from a normalized number, such as 10 lawyers per billion of revenue. The bigger the law department, the larger the absolute numbers, whether of lawyers, internal spend, external spend, cases, office locations, library books or whatever. Small legal departments just don't match up. However, if you divide the absolute numbers of departments of all sizes by revenue (or by employees, market capitalization, years in business, depreciation on hardware or some other figure they all have), departments of every size can compare themselves to each other on a similar quantitative footing. Both the top number (the numerator) and the bottom number it is divided by (the denominator) tend to grow or diminish in proportion. This technique of normalization allows benchmarking analysts to include law departments of vastly different sizes and yet produce their metrics on a comparable basis.

Let's clarify this with some examples. A law department with three lawyers has 60 cases pending. Another company has 20 lawyers and 200 cases pending. The absolute numbers are much bigger, obviously, for the second department, but is the ratio of lawyers to cases? Without normalizing those absolute numbers, it is difficult to say which law department has a higher caseload for its attorneys. So, you divide each department's cases by its lawyers. The first has 20 cases per lawyer whereas the larger department has only 10 cases per lawyer. Once you normalize the two departments - adjust their figures by numbers of lawyers, you can speak about them in the same proportionate terms and clearly describe differences in caseload.

Consider another example based on compensation. If you know that lawyer Alpha makes \$200,000 in base and lawyer Beta makes \$220,000, all you can say is that Alpha makes \$20,000 less than Beta. If, however, you were to divide each lawyer's salary by the lawyer's years out of law school, you will find that lawyer Alpha - who graduate five years after Beta - makes considerably more in terms of dollars per post-law-school years. Alpha at 15 years makes \$13,333 per post-grad year whereas older Beta at 20 years from law school makes \$11,000 per year. That calculation normalizes the two lawyers' pay by years of legal experience and shows that Alpha is doing considerably better.

Whenever you normalize a set of numbers by dividing each one in the set against the same other number (which may well vary from department to department, such as numbers of cases pending or years out of law school), you can also reverse the division - make the former numerator the denominator - and produce a different version. For example with a group of law departments you could normalize the patent annuities paid each year by dividing those payments by the number of their patent records. Alternatively, with a mathematical flip, you can divide the number of patent records by the annuity payments per year. Thus, one company with 100 patents around the world and payments of \$100,000 yields a normalized metric of either \$1,000 per patent record or one/tenth of a percent of a patent record per dollar of annuities (but you probably multiply both numbers by one thousand and express the ratio as one patent record for every thousand dollars of annuities).

Which fraction to use depends mostly on familiarity. General counsel may be more familiar with one way to express it, such as total legal spending as a percentage of revenue, but it is just as accurate to describe the normalized result as revenue dollars per dollar of total legal spending.

### Rees Morrison

Rees Morrison, Esq. is a partner at Altman Weil, Inc. with countless interests in legal data analytics. He is also the founder of General Counsel Metrics, LLC....

Bio and more articles