I’ve been catching up on podcasts, and a few weeks ago I listened to episode 673 of 99% Invisible - The Score. Most of the discussion focuses on how metrics have become prevalent in our modern life, and how maximising a particular metric is often not taking us to the goal that we had intended. These thoughts resonated with me, as my perception from my working career is that metrics usually have a whole host of unintended consequences, and often seem to do more damage than good.

Goodhart’s Law has always seemed to be correct in my working experience. Effectively it states that when you use a metric as a target or goal, then people will game the metric and it ceases to be useful for its original purpose of understanding how the business is performing. This seems to me to be particularly the case when it comes to metrics that attempt to quantify the performance and contribution at an individual level.

A good example from my own experience is a metric related to the total dollar value of your sales pipeline. When this is set as a target, it is easy for people to inflate the metric by including lower likelihood sales opportunities and inflating the value of the opportunities in the pipeline. The metric doesn’t improve the sales outcomes, and now you are chasing down a different problem of trying to understand why the conversion rate from the pipeline to sales seems to be lower than it previously was.

Some organistions seek to compensate for this by adding more and more metrics. I think that this just adds additional administration and further dilutes focus away from business performance. When you have so many metrics to monitor, you can’t put as much focus into the ones that matter.

A further challenge is that in the sorts of organisations I have worked with and worked for, each employee often has very different skills and contributions that they bring to the business. Some will be excellent project managers and great at following processes and procedures. Others are great at lateral thinking and problem solving. Others are excellent communicators. While there is often a requirement for a common baseline of skills, the business succeeds best through bringing these disparate talents together and applying them to where their skills are most effective.

Unfortunately, the performance metrics that I have seen are often standardised across individuals and fail to recognise the different ways that they bring value to the organisation.

This also ignores the fact that the way a person contributes value to an organisation across a performance year will often change in response to the business environment. They may not measure up against the metrics that measure how you thought they would contribute to the business, but that will probably be because they have change their focus in response to changes in the business environment.

There is a lure to the idea of metrics as providing an objective way of comparing people’s performance. In practice there is subjectivity in how the metrics are chosen and often in how they are evaluated. Further, just because something is easy to measure, doesn’t mean that the it is the most important factor.

Having said this, businesses need to make money to survive, provide a return for their shareholders, provide income for their employees and provide needed services within our economy. Some metrics are important. In my opinion, the classic accounting metrics such as revenue, expenses, profit, etc are important. However we need to remember when interpreting these, that even these metrics can be gamed in ways that are not actually valuable - e.g. radical cost cutting or increases in leverage, that can inflate value in a given year, but result in poor performance a few years down the track.

Paul Graham, has a contrary view that he expresses in How to Earn a Billion Dollars. He argues that obsessive focus on growth is a key requirement for successful start up founders. He’s a smart guy with a serious track record in this space, so his thoughts are worth considering.

I’d also note my own inherent biases. I’ve seen lots of examples of how metrics fail, or seem to lead to suboptimal outcomes, but perhaps things are even worse if you don’t have these metrics?