The adage, you can’t fix what you can’t measure, gets repeated ad nauseam within most companies. Metrics are found at the heart most continuous improvement initiatives. However, despite the gospel, this understanding of metrics, they are often poorly presented and do not help drive improvement. Here is a common list of sins:
- Too Many Metrics
- Not Intuitive
- Poorly Labeled
- Inappropriate Graph and Scale
- Out of date
- No Benchmark
And there are many more. These are some elements that I look for when I review a metric.
Tells A Story
This doesn’t mean the metric is biased to support a position, but rather it conveys information that allows the reader to understand the issue being presented. Really good metrics can help people postulate potential root causes and develop action plans.
Ask someone unfamiliar with your business to read the metric and explain what it is reporting. Make sure the information is labeled clearly. Label metrics with simple English; try to avoid abbreviations and jargon. Do not overcrowd the data; make it visually pleasing. Use some imagination in how the data is displayed. Make the metric a compelling read.
The reader wants to know why they should care about this metric? Think about a good newspaper article, it all starts with a catchy headline and a lead sentence. The metric should state what it indicates (primary factors and secondary factors). For example, we may track on-time delivery because it relates strongly to customer satisfaction. Make sure the rationale is stated on the metric. Some metrics may have correlations in multiple areas. You may want to indicate secondary issues that the metric correlates to as well.
Having too many metrics overwhelms people. Be selective, display the ones that are relevant for the most important issues facing the organization. You may want to track more, but only display the ones that are essential.
Make sure the metric being tracked can be influenced by the actions of the group. If the metric is not actionable, then we might as well track the weather.
A metric should show the history of performance over a reasonable time frame (months, quarters, etc). Sometimes the comparison needs to be with a corresponding period, rather than an adjacent period. For example, comparing 3rd quarter performance this year to the 3rd quarter of the previous year may be more telling than comparing the data to the previous quarter. This is especially relevant for businesses with a strong seasonal component.
Metrics should include enough subsequent periods to indicate a trend. They don’t always have to start at the beginning of the year and end in December. Give some thought to what period of data makes sense. Sometimes a year of quarterly data makes a better story, sometimes the last 6-months will make more sense, A rolling average line can help readers spot trends in spiky data. Your choice of period for comparison can make or break the effectiveness of your metric.
Establish a benchmark for excellence based on the type and size of organization being measured. This is not an easy task. It takes time and research. In the end, if you have no other data points, make your best guess based on the experiences of the people in the room. Avoid using general data that may not be appropriate. What is world-class performance for Intel may not make sense for your business. Write down how the benchmark was developed, otherwise people forget how it was derived.
What is the goal for this metric at the start of the period? This could be set at the beginning of the week, month, year etc. This should be visible on the metric being presented. This tells the reader how well the organization is performing to their expectation. Remember Tampa Bay was not built in a day; the team's goal for a metric may not be the world-class benchmark set by Toyota in 2009.
Action and Status
What is the current condition of the metric being tracked (i.e. Red, Yellow, Green)? What problems are preventing the group from achieving a higher level of performance? What actions are being taken to improve the issues being measured? When does the team expect these actions to be completed? Your metric should allow the reader to understand how the organization is responding to the data.
The trick is to do this in such a way as to not overwhelm the reader. Be economical in addressing this need. For example, root causes need to be only a few words. If the team can develop codes for root causes that enable them to be used with other metrics, this helps identify significant opportunities. Use common colors and shapes to link appropriate data and make the information more intuitive.
Change the metrics from time to time. There will be some you always want to display, but others can be rotated. Without change, information fades into the background. That's why they have a big basement at the Smithsonian.
This is true both in terms of accuracy and performance. Getting the data exactly right for a metric can be a time consuming task. If a metric takes too much time to complete, it will soon be headed to the metric graveyard. Close counts in many cases. Don't let a pathological need for accuracy get in the way of good information, leave that for the accountants.
The other element of good enough relates to performance. Pushing too hard on a single metric can have a negative impact in other areas of the business. Don't let the crowd or the boss bully you into driving for 99.8% for on-time delivery if increasing the performance in this area is no longer the key element to improving customer satisfaction. Everything comes with a cost; remind the group of this when the overachievers get out of control.
Well done metrics can have a powerful effect improving a company’s performance. They take time, creativity, prototyping and feedback to get right, but it’s worth the effort. In the coming weeks, I will explore other aspects of metrics and their use. What have you seen both good and bad when it comes to metrics? Send me some examples. Helping companies measure their performance more effectively is a key element to sustaining their competitive advantage.
About the Author
Tom Petersen is the Managing Partner of ThreeCore, an operational consulting firm headquartered in Beverly, Massachusetts. Tom consults for multinational companies engaged in the design and manufacture of high-tech products. His team is dedicated to helping companies create competitive advantages using innovative strategies and process-driven improvement. For more information go to http://www.threecore.com or follow Tom on twitter @3CoreConsulting.