We often talk about finding the right KPI. The perfect one. The golden one. The number that tells us whether things are going well and answers the questions everyone in the organisation might have. It sounds appealing. But the more I work with measurement, the less I believe that this number exists.

Not because KPIs are a bad idea. Quite the opposite. A good KPI can create a common language across an organisation and help people focus on what actually matters.

The problem is that we sometimes start looking for the KPI before we have properly understood who needs it, and why.

There isn't one set of questions

An organisation doesn't have one single end user.

A marketing specialist might want to know what happened yesterday. A product manager might be looking at what has changed over the past few weeks. A commercial team might care about whether a particular activity is contributing to revenue. And when leadership meets to discuss the quarter, the questions are likely to be different again.

They are all looking at the same business. They just aren't asking the same questions. And that means they shouldn't necessarily be looking at exactly the same measurement.


This is why I like thinking about measurement as a hierarchy.
At the bottom, we have metrics: individual measurements that describe something that happened. We can combine metrics into indicators that give us a more meaningful view of a particular area of the business. And then, from those indicators, we can define KPIs - the Key Performance Indicators that the organisation agrees are important enough to share and follow consistently.

The hierarchy itself is relatively easy to understand. The difficult part is deciding what belongs in it.

The problem with KPI workshops

This is where KPI workshops often become surprisingly difficult.

I've been in - and seen - many workshops where people are asked what the organisation should measure. Someone presents a proposed set of KPIs. People look at them. There is some discussion. And then, quite often, people agree.

Sometimes because the KPIs make sense. Sometimes because they already exist. Sometimes because they were proposed by leadership. And sometimes, I suspect, because it is surprisingly difficult to come up with something completely different when you are sitting in a meeting room being asked to define the important KPIs.


There is also a practical problem. If the organisation already has a set of KPIs, changing them can be uncomfortable. People have been using them for years. Reports have been built around them. Targets have been attached to them. Management meetings refer to them. Even when everyone agrees that the KPIs aren't perfect anymore, replacing them isn't a small decision.

And if there are no established KPIs, starting from a blank sheet can be even harder.

So I like to approach the problem from a slightly different direction. I call it the Santa Claus list.

Think like a child with a toy catalogue

Think back to when you were a child and you received a toy catalogue before Christmas. You went through the pages and circled everything you wanted. You didn't stop to calculate whether it was affordable. You didn't wonder whether it would fit in your bedroom. You didn't ask whether your parents would be able to find it in the shops.

You simply thought: I want that. And that. And definitely that.

That is the mindset I want to recreate in the first part of a KPI conversation.
I ask people to come without their computers. No dashboards. No spreadsheets. No existing reports open on the screen. Just a pen and a piece of paper.

And then I ask them to imagine that there are no technical constraints. If you could have anything you wanted in your measurement environment, what would you want to see? What would help you understand your work better? What would help you make a better decision tomorrow? What would you want to see every morning? What would you want to understand every week? And what would you want to know when you sit down at the end of the quarter and ask yourself whether things are actually going in the right direction?

At this point, I don't want to hear whether it is possible. I don't want to hear that the data isn't currently available. I don't want to hear that the tracking would be complicated. And I don't want to hear that "we have always used this KPI".

For a moment, everything is possible. This is the Santa Claus list.

Separating what matters from what we can measure

There is a reason I think this exercise is useful. We often let the measurement environment define what we consider important. We look at the data we already collect and ask what we can build from it. We look at the existing dashboard and decide what should be on the next version. We look at what the analytics platform can report and build our KPIs around those capabilities.

But that reverses the problem. The fact that something is easy to measure doesn't mean that it is important. And the fact that something is difficult to measure doesn't mean that it isn't.

If someone says, "I really want to know whether our most valuable customers are becoming more engaged over time," the first question shouldn't be whether that metric already exists in the data warehouse. The first question should be: is this something that would help this person make a better decision?

If the answer is yes, then we have something worth investigating. Maybe we can already measure it. Maybe we need to combine several existing metrics. Maybe we need better instrumentation. Maybe we need data from another system. Maybe we discover that we don't have a sufficiently good definition of "valuable customer" yet. And maybe, eventually, we discover that the measurement would be disproportionately expensive compared with the value of the decision it supports.

All of those are useful conclusions. But they are conclusions we can only reach after we understand what matters.

The technical conversation should come later

This is why I deliberately separate the first conversation from the technical reality. The Santa Claus list is not the final KPI framework. It is not even necessarily a list of things that will ever appear on a dashboard. It is a way of discovering what people actually wish they could understand about the business.

Once we have that, we can start bringing reality back into the room. We can look at what is already measured. We can question definitions. We can compare different people's lists. We can find common themes. We can identify where several teams are actually asking the same question using different words. And then we can start discussing what is technically possible and what is worth implementing.

Sometimes the answer will be a simple metric. Sometimes it will be an indicator built from several metrics. Sometimes it will become a KPI that everyone can share. And sometimes the answer will be: we don't need to measure this after all. That is also a good outcome.

Maybe the golden KPI isn't the point

I don't think the goal of measurement should be to find the one number that everyone agrees is important. The goal is to create a measurement system that helps people make better decisions.

That means accepting that different people will need different levels of information. The person working with a channel every day may need detailed metrics. Their manager may need a smaller set of indicators that puts those metrics into context. Leadership may need a handful of KPIs that provide a consistent view of performance over a longer period.

These aren't competing views. They are different levels of the same measurement system.

And perhaps this is why I am increasingly reluctant to ask: "What is our KPI?" I would rather ask: "What decisions do we need to make, and what would we need to know to make them better?", "What is the problem we are trying to solve here?" The KPI can come afterwards.

So before opening the dashboard, before checking what your analytics platform can provide, and before deciding whether the KPI you already have is still the right one, try something simpler. Give everyone a piece of paper. Give them permission to forget the technical constraints for a moment. And ask them to make their Santa Claus list.

You might be surprised by what they actually want.