Designing KPIs that Work– Do the Math
By Dr Sara Cullen | March 3, 2016
It is difficult to develop KPIs (Key Performance Indicators) that work well in practice. Unless you design
KPIs carefully, they will have varying degrees of inaccuracy and incompleteness, or get implemented in ways that ‘get the numbers’ but not the results.
Inevitably, you will lose confidence in KPIs as a performance management tool, and possibly lose confidence in the provider. This is an easily avoidable situation. Design KPIs that work – do the math.
The process of developing meaningful performance measurements is an intensive one. The method itself is not especially intricate, but depending upon the complexity of the work under contract, KPIs can take many iterations until the measurements are bulletproof (that is, unable to be interpreted in any other way except that which you intend).
Most contracts stop at only specifying the KPIs, as shown in Table 1. While at first glance, this may appear satisfactory, it is not.
Table 1: Typical specification of a KPI
|Connections||95% of Authorized Services connected within 1 Business Day||100% of Authorized Services connected within 1 Business Day|
If the math underpinning the KPI is not specified, then a provider self-reporting its performance (which is the norm) is allowed discretion as to how to calculate that performance. Not only can the provider make up the formulas, they also get to make up which sources of data they want to use. Unsurprisingly, the KPI report will typically show a good result. No sensible provider left with total discretion as to how to report its performance would do otherwise.Table 1 is inadequate because it lacks a specification explaining how the KPIs must be calculated. Most disputes about KPIs centre on this one root cause. It is not the numbers that are most argued about, it is how the party doing the calculation – the math – determines it, and when and from what data.
A better example of a KPI is shown in Table 2, which includes the specification for the calculation, and its timing, formula and source data.
Table 2: Better specification of a KPI
(expressed as a %)
|Connection turnaround time||95% of Authorized Services connected within 1 Business Day of Connection Request||100% of Authorized Services connected within 1 Business Day of Connection Request||Monthly||# Authorized Services connected within 1 Business Day / # Connection Requests issued during the month||Connection Request Log|
|100% of Authorized Services connected within 2 Business Days of Connection Request||# Authorized Services connected within 2 Business Days / # Connection Requests issued during the month|
If you do not know how to do the math, you are not ready to have a KPI that works. Go find out – do not put out a tender, and never sign an agreement with this kind of detail missing.
Getting the math right is only one of seven key steps in developing robust KPIs that we go through in the Contract Scopes, SoWs & KPIs Masterclass course that’s running 28-29 April in Sydney.
But the KPI problem it is quite a basic one, and one that is too often the cause of easily avoided trouble. Take the time and the care – for quite a minor investment in conscientious thought, it will pay off immensely.