Proactive metrics vs. predictable process

In a discussion with a client this week, we compared a need to find proactive metrics and what it means when a reactive metric is predictable.

A proactive metric would be one which provides an early indication of the final quality or performance of an item or service.  These are difficult to identify, but they are typically input values to a process.  If you have a good understanding or model of the Y=f(x) relationship, the prediction of the output can be reasonable.

For a predictable process, even if you are evaluating the final quality or performance of an item or service, is not quite “proactive” but it provides a possibly better view of the performance.  Predictable indicates that the process output is consistent or stable as measured by a control chart.  In this case, you can predict the same average performance, as long as the process is not changed and the variability of the inputs stays about the same.

Both provide indications of future performance; both depend on the process not changing.  So which is better?  I believe that any process output measured with a control chart is effectively a proactive metric, if you are evaluating the predictability and not just the capability.