Friday, December 14, 2012

Metrics, metrics

Metric: a standard of measurement -- Example: no metric exists that can be applied directly to happiness — Scientific Monthly

Measure: the dimensions, capacity, or amount of something ascertained by measuring

At Leading Strategic Initiatives, Greg Githens has a worthy posting on what makes a good metric. Githens offers six qualities for metrics, abridged and repeated here:

  1. It measures something important. ... metrics reflect the imperatives of the individual or the organization.
  2. It has relevance to the audience. Since ... initiatives have difference stakeholders, one of the biggest challenges is to prioritize the audience and tailor [metrics for]... them.
  3. It measures something that is directly controllable by individuals or small groups. This suggests that metrics are local, and connected to action.
  4. It is resistant to gaming. .... the metric is difficult for self-centered actors to manipulate.
  5. It is a member of a very small, lean set of measurements. Since people have a limited span of attention, we want to keep the metrics to a handful
  6. [It is a member of a] ... set of metrics [that] includes both leading and lagging indicators. No one drives their car by focusing [exclusively] on the rear view mirror, they [also] look down the road to see the turns and respond to the threats.
If you're thinking about applying these ideas and are looking for some specific project metrics, do a read of John M. Green's essay at entitled "System Measures of Effectiveness". In the abstract, Green writes:
Proper selection of performance measurement attributes is essential to [the performance analysis] process. These measurement attributes commonly called measures of effectiveness, or MOEs, provide quantifiable benchmarks against which the system concept and implementation can be compared.

In this paper, you'll find this table of metric qualities (charateristics), very similar to what Githens posits: