Oops! The input is malformed!
Originally published abril 29, 2008
Ahhh, springtime! It is a beautiful time of year. The flowers are starting to bloom. People are starting to emerge from their “winter doldrums.” The upsets in the men’s division of the NCAA Tournament again showed us that anything can happen. On the women’s side, Pat Summitt and the University of Tennessee showed that there are a couple of things that you can count on: death, taxes (speaking of springtime…) and Tennessee playing in the Sweet 16, if not the title game.
Another thing that you can count on is that in the world of business intelligence (BI), the words of former UCLA basketball coach John Wooden will ring true:
“Never mistake activity for achievement.”
Wooden used this concept to show that an individual or organization should not see the simple act of showing up or going through the motions as achievement. Rather, they should focus their efforts on the actions that produce significant, or strategic, results for the organization.
With the introduction of operational business intelligence to match with the existing practice of traditional enterprise business intelligence, there has been a flood of metrics to measure the “achievements” of the operational business as well as the enterprise business. Most people call these metrics key performance indicators (KPIs). However, if everything is a KPI, aren’t we abusing the “key” concept in the KPI?
While I might be abusing the words of Soren Kierkegaard, I fully believe that most people wander around labeling just about anything that can be measured as a KPI. In this discussion, the first aspect of KPIs that should be addressed is “What is a key performance indicator?”
With a little help from Wayne Eckerson, I have the following “textbook” definition of the progression of the difference between a metric and a KPI:
Most people do not have that definition in their back pockets. They assume that if something can be measured or trended or predicted, they can put the label of KPI on it. In this, we start to get into the “activity” side of KPI definition and development. Just because you can measure something does not make it a KPI.
A more strategic approach should be used similar to the approach advocated by Neal Williams, Founder and Chief Scientist of Corda Technologies:
“We’ve found that the best way for companies to define KPIs is to hone in on what metrics best reflect strategic and operational performance, whether it currently exists as a data source or not. By employing a top-down approach, defining what metrics are truly key performance indicators and what are just noise also helps identify where there are holes in an organization’s tracking and monitoring systems.”
Air temperature associated with the weather is an interesting example. You can measure it. You can put a trend on it. I would daresay that you can do a reasonable job of predicting its behavior (or at least that’s what most meteorologists claim…). However, the part that is usually missing is the context surrounding the measure that would turn it into a KPI. What would make it a strategic or operational driver for an enterprise?
In the world of information technology and business, you would be hard pressed to find someone who finds air temperature to be an important enough measure to make it a KPI … that is, until the air conditioning fails in the data center or the office. At around 80 to 85 degrees Fahrenheit, both servers and people do their jobs less efficiently and start to complain. At that point in time, everyone can agree that air temperature could be an important metric and probably rises to the level of a KPI that can impact the strategic performance of an enterprise.
I do not advocate that air temperature become a strategic KPI (unless, of course, you are an airline and you are measuring the amount of weight that a plane in Denver can hold and still take off on a 100 degree July afternoon). However, air temperature has its place within the catalog of metrics and potential operational KPIs. As such, it should not be forgotten after the air conditioning is repaired and the context of its use is past. This does not mean that this metric should be lost, but rather that it should be “archived” and used later when certain issues will again bring that metric to the forefront.
The words of the Roman philosopher Seneca lead to another aspect to metrics and KPIs that has arisen along with operational business intelligence. A metric that is important to one organization’s daily operations or process can have no bearing on another part of the organization. However, taken in another light, the operational measures of one organization can contribute to the financial KPIs of another and/or the entire organization.
However, most departments are loath to share their information with other parts of the company unless requested; and even then, they might get into a positional negotiating stance that this data and the associated metrics “belong” to them. This gets to the point of an excellent article by Len Silverston about the difference between data “mine-ing” and data “ours-ing.” When you keep your information to yourself, you often get into situations where the company suffers.
This silo-ed approach to measures, or rather the lack of a comprehensive look at metrics and measures across the organization, can be a major stumbling block to organizations finding that “magic combination” of KPIs that truly give the company visibility into how the company works and how to improve that performance.
The telecommunications industry might be the poster child for issues associated with mis-definition of measures and KPIs, and the ineffective use of those measures across departmental lines.
Last year, I wrote about a study from Nathaniel Palmer that talked about how telecommunication service providers are notoriously adept at creating new initiatives in individual departments or regions and that those efforts aren’t coordinated across the enterprise. I have not seen much this last year to indicate that these issues are going away.
Where the business intelligence/data warehousing (DW) organization comes into play for these issues is more a political than a technical issue. It’s how to get the various organizations to share their information.
Part of it comes from the leadership of data steward. This data steward should not only be the “marketer” for the measures/KPIs that are produced by the BI/DW organization, but he or she should also be on the lookout for the measures/KPIs that other parts of the organization are creating on their own, either from an operational perspective or a financial perspective. Not that the data steward should be looking to “take” those measures/KPIs into the BI/DW organization; but if the data steward created an active catalog of measures/KPIs being used in the organization, perhaps the data ”mine-ing” would slowly come to an end.
Recent articles by John Myers