Proxy measures do not allow us to understand what performance is doing. To understand what performance is doing in our organisation we need to work out what to measure, then how to measure it.
The classic idea of a ‘proxy measure’ is an indirect measure of an outcome, particularly where the data is hard to come by. For example, ‘infant mortality rates’ could be a direct measure of the efficacy of a community healthcare system, but is often used as a proxy measure for the social/economic community welfare (source: https://govex.jhu.edu/wiki/proxy-measure/).
The proxy measure is referred to as an indirect measure, but something that can be correlated to a related outcome. For example, at a community level Infant mortality rates would be a direct measure of healthcare quality – but it also could be a proxy for the economic and social welfare of a community. Similarly, unemployment rate is a direct measure of the level of unemployment, or the level of engagement in the national workforce. but unemployment rate could also be seen as a proxy for the overall state of our economy. Perhaps proxy measures like this can provide some feedback on outcomes at a community, or national level, as there are many systems that contribute to those type of outcomes.
Proxy measures might be okay to find some correlation, but with understanding and measuring organisational performance we need to understand causation.
It seems many organisations are taking this idea of a proxy measure and applying it generally to organisational performance. Often where it seems ‘hard’ to find the optimal measures, we see the solution as selecting measures where the data is readily available. Even though we know these measures aren’t really providing good feedback about the results or outcome we are trying to achieve, these measures are then said to be ‘proxy’ performance measures.
Whilst this might seem like a reasonable shortcut at the time, these types of proxy measures cost time and money but they add very little if any value. The cost in time and money come in the effort to collect the data to produce the measure, to report the proxy and then in all the lost time in meetings debating how to interpret the measure and debates on what it is telling us.
To me this short-cut and doesn’t help. Picking available data and reporting it, is not measurement of organisational performance. We need to know what we are first trying to measure, then design and select the most meaningful measures of performance.
Instead, we could use a proven method (such as PuMP). Within this approach we would first define the results/outcome we want, then design the performance measures that will give us the best feedback on performance. What we are looking for when trying to understand what performance is doing
Step two of the eights steps within the PuMP Blueprint is exactly about “what” to measure. Too often when we think of broad terms, we jump straight to figuring out how to measure it, get stuck and look for shortcuts. Before we even start talking about performance measures, we need to be able to articulate what it is we want to measure. We need to define the results we seek in a language that makes them measurable.
Once we have a result that is measurable – then we can design and select performance measures that will tell us how are going in trying to get to that result. Step three of PuMP uses a Measures Design technique that guides us through a process to ensure we have meaningful measures for that result.
Proxy measures do not allow us to understand what performance is doing. No more proxy measures for performance – make to effort to make what you want to measure, measurable, then design measures that will provide feedback on the result. If the result is important, the investment in effort and time will be worth it.