Putting Your Scorecard to Work, Pt. 3

Analytical Performance: A Step-by-Step Guide

This is Part 3 and the last post in a discussion of how to make corporate performance reporting more analytical. If a multivariate analytical reporting model is the ultimate in corporate performance management processes, what, then, are the stages by which organizations would achieve that goal? Of course, all such maturity models have to have five stages, so I will stick to that convention.

The first stage is simply to produce accurate and timely financial reports. I view this as a low goal to which no one should aspire, but so far this level of performance management is all that is required by regulatory bodies (e.g., the SEC in the United States). However, the number of companies having to restate earnings or delay financial reports suggests that even this stage is not yet a reality for many firms.

Stage 2 involves successful implementation of the balanced scorecard. Financial and nonfinancial indicators of performance are displayed on one page. Most balanced scorecards adhere to the Kaplan and Norton recommendations, and report on the categories of “financial,” “customers,” “internal business processes,” and “learning and growth.” Instead of simply taking their advice, your organization should decide what categories make sense for your strategy.

Stage 3 begins to relate the relevant measures of nonfinancial performance to financial performance at the level of logic and common sense, but not the quantitative level. For example, stating that better employee satisfaction should yield better customer service, and that should yield higher sales, is a form of a strategy map. Kaplan and Norton have called such a set of relationships a “strategy map.” (http://en.wikipedia.org/wiki/Strategy_map) While it is useful to know an organization’s logic about what drives performance, the logic could, of course, be incorrect if not validated statistically.

Stage 4 is the model I have described in previous posts in which nonfinancial and financial variables are related to each other in a statistical fashion. Companies have made only limited steps toward this state, and no organizations we know of can be said to be fully at Stage 4. However, the examples in the previous post clearly suggest that companies are moving in this direction. The actual statistical models to accomplish this step are not very difficult—even I could do them—although measuring and collecting the data for the model would be. The model would need to be refined, perhaps multiple times. Ideally, the coefficients in the model would correspond closely to management’s ideas about what drives the business—although there are likely to be some surprises. My sense after researching the issue is that it’s more likely that this works in an organization that can measure performance at the branch, store, or small business unit level. At the level of a large corporation, too many of the variables are difficult to measure with precision.

Stage 5—corporate performance nirvana—would require not only having a quantitative model in place, but also embedding it into the minds and performance evaluations of employees. If they know what really drives performance, they can emphasize and be measured on those factors. When Sears (http://hbswk.hbs.edu/archive/801.html) adopted a variation of the service profit chain in the 1990s (for some reason they abandoned it later, even though it led to great financial performance improvements), they felt that this stage was the most difficult to achieve.

I am sure that organizations will eventually move in this direction, because the data are increasingly available in scorecards, and because the value of analytical performance reporting is obviously high. But I confess to some perplexity about why it’s taking so long!