For over 25 years I have managed analytical teams for growing agencies, large agencies and for large clients. Before the term “KPIs” was introduced, we just called them performance metrics. Then KPI became in vogue and that’s when the fun began. It seemed everybody wants to define, by committee, the single most important KPI, a “Supreme or Main KPI”, for the business before they set off experimenting ways to improve something. This often led to endless debates that stalled action, like the metaphorical question of how many angels can dance on the head of a pin.

My advice? Just start doing. Yes, identify your main KPI. Chances are you might not directly affect it, but you can influence the smaller KPIs that contribute to the "Supreme KPI." Begin with those.

Take, for instance, a baseball team. Winning the game is the "Supreme KPI." As a hitting coach, I can influence that KPI, but I can’t directly cause it. Instead, the KPIs I can directly influence a team win are things like improving batting averages, on base averages, walk and strikeout percentage, improving the overall productivity of the lineup order, etc. These are specific areas I can influence that helps the team win.

Find your main KPI and the smaller, controllable "sub-KPIs" that you can optimize to improve the overall KPI. Building an environment that focuses on testing and learning will help continuously enhance these sub-KPIs, which in turn, supports the main KPI. Think of it as a cycle of constant improvement, with KPIs acting as the measure of success in your experiments.

The key is to adopt an experimentation mindset. Make your first move with KPI’s you can directly impact.