Non-financial metrics or indicators can provide management with additional insights into the health, performance and capacity of a business process or system. Organizations may use these to improve budget predictions, augment a 'Balanced Scorecard' framework or help manage an outsourcing program. They can be powerful problem-solving tools. In general, there are four major reasons as to why organizations would develop a metrics program:

• Characterize a process or system.
• Evaluate a process or system against standards, benchmarks or history.
• Improve processes.
• Predict future outcomes.

Characterization: Establishing a set of simple measurements or indicators can help understand the current state or behavior of a business function or system -- current computer utilization or call center staff workload, for example. Metrics can provide a vital tool for communicating an invisible reality to other business areas. This is always the first step for more sophisticated applications.

Evaluation: Measurements or indicators would be used to manage the business process. Report production times, error rates, enqueued service call counts are examples of measurements that could be used on an ongoing basis to manage a process. Evaluative measures may also be compared to standard or benchmark information in an operational framework. One common approach is to calculate ratios around relationships between support staff, users and equipment -- this may be useful for comparing organizational efficiency. But approach to change management and requirements volatility must be factored in. Service level agreements often contain standards of performance that ongoing operations must manage to. Both measurement methods and governance processes need to be accommodated in the SLA for this to work, however.  

Improvement: Measurements can be used to guide or manage a process improvement program and show when objectives are achieved. Technical indicators are often much more readily accepted than indicators of staff behavior - but many process improvement projects exist to and are measured by changes in staff behavior reflected in costs and service times. It is essential that staff be aware and supportive of the process and chosen metrics at the onset.

Predictive: Predicting future costs, system performance, storage requirements or staffing are common motivations for establishing a metrics program. Effective predictions requires an understanding of the relationships between the ultimate business or technical drivers and the predicted characteristics. Models can encapsulate this understanding, but it is prudent to validate against historic information. This approach to process metrics is often referred to as 'activity-based costing' and is ultimately derived from industrial engineering approaches to motion and time measurement -- but applied on a large scale.