Many have forgotten that the earliest data warehouses were built to provide reports to management on business performance. Early widespread definitions of data warehousing actually said data warehouses were "to support management needs." This was a good start, but the untapped potential benefits to a larger audience awaited. The later application of data warehousing to the larger masses of knowledge-workers and, more recently, everyone, is proof of that.

Nonetheless, data warehouse applications as well as complete, standalone data warehouse efforts exist today for business performance needs. These management business performance needs tend to get lumped under key performance indicators (KPIs). Many use summarized versions of the larger body of data warehouse data that everyone uses, with a specialized application interface on top.

Using the same data that serves many other purposes in the organization as a basis is actually part of forging great organizational synergy and is much more effective for facilitating communication. Under the notion that "what you don't measure, you don't improve," these KPIs need to be measured with data that is clean, updated, shared, integrated, has a historical perspective and has widespread buy-in to its build process. That's the data in the data warehouse.

KPIs are usually measured top down; but if the supply chain manager has access to the same KPIs as the CEO ­ and knows which ones she is responsible for ­ she can track her own performance and adjust accordingly. Too often, however, different business units are running off KPIs measured with different sets of data. When these KPIs are taken in the aggregate, they do not support any sound business strategy, much less the one the CEO would like to promote.

What KPIs should be measured? Depending on whom you talk to, several hundred could be suggested. However, I've found that the most effective KPI-using organizations track back from company profits to the enabling factors specific to the current picture of income and expenses. Several hundred KPIs are not necessary. Even companies in the same industry do not need the same metrics. Many efforts fail when a flood of information at the starting point drowns the executives.

Some areas that should receive consideration are:

  • Procurement ­ Fill rates, on-time performance metrics, costs, supplier management, quality management.
  • Supply Chain ­ Cycle times, supply chain costs, transportation costs, capacity, adherence.
  • Customer-Facing Metrics ­ Channel utilization, payment preferences, contact center utilization, fraud, returns, customer satisfaction, warranty work, incomplete orders.
  • Productivity ­ Inventory levels, wait times, employee productivity, day's sales outstanding, labor costs, yield.

Start with measures directly interesting to the bottom line and grow from there in an iterative manner, using interesting metrics that you are able to collect from the data warehouse.
A major shift in data warehousing in the past few years in support of management needs is the ability to drill down to detail. Each KPI needs to be supported by detailed data that is "drillable" from the KPI application. Metrics need not drill to a single summarized data point. Rather, the most effective ones actually comprise several metrics prorated into the overall metric. Customer satisfaction is a good example of this.

Overall, customer satisfaction is deduced from several data points such as direct customer surveys, churn and retention, share of wallet, etc. ­ each a valid metric in its own right but interesting primarily as an indication of customer satisfaction. At some level, the red-yellow-green of "customer satisfaction" is the interesting data point.

The real-time nature of the metrics is also possible with data warehousing. Regenerated by triggers or processes at the end of batch ETL, the metrics are never too far removed from a currency that is meaningful. With the real-time nature of many data warehouses today, KPIs can be current up to the moment. In high-volume environments, 3-D data visualization can help sort through the volume of data in real time.

Finally, KPIs need not be simply "scored," but also full cause-and-effect relationships need to be generated. More advanced KPI applications generate not only the root causes of the score, but also early indications of future metric levels.

If you are not presently using your data warehouse for KPIs, consider it. You probably have enough data to begin. By summarizing and focusing the detailed data you deal with every day into something coherent for management, you help ensure synergy and attainment of corporate goals. By using data warehouse data to support real-time KPI metrics generation, you can very effectively score and support strategic business objectives, monitor progress in real time, provide drill to detail capability, generate cause-and- effect models and, most importantly, measurably improve the bottom line.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access