Goals and KPIs, and Effective Managers

Effective managers love the concept of a KPI (key performance indicator a lot) 💘. These people come to a company, think through business processes, form hierarchies, appoint mid-level managers, implement a business KPI or two, and start measuring everything—honoring the memory of good old Drucker. It looks like they’re doing a great job, completely by the book. They follow the prescriptions from the manuals they’ve read, the advice they’ve got, great KPI examples, and their short consulting experience in five or so companies. Basically, everything should work just fine—but usually, none of this leads to a desired outcome.

A KPI almost always becomes a business goal. Why? Well, as soon as you start measuring something, you immediately feel the urge to set a measurable goal for it, too. When you start measuring the company’s revenue, the financial KPI tells you: “next year, our revenue should grow by 30% to two million Canadian pounds.” When you start measuring the churn or bounce rate, the desire sneaks in to decrease the churn or bounce rate from 15% to 8%. You see where it’s going, don’t you? When you start measuring the number of bugs in production, the VP of Development will declare that “we need to have absolutely no bugs in production in the next quarter.” So-called smart KPI pins you down to chasing some “success,” understood as bare numbers.

But hey, why is turning KPIs into targets doesn’t lead to the desired outcome? The problem is that the entire team will rush to achieve that very specific goal by any means. And this is how people will try to increase the turnover by selling to non-target customers—resulting in skyrocketing churn rates. To fix churns, 70% of the customers can be given a 90% discount, killing the whole unit’s economy. To get rid of bugs in production, a testing phase for three months could be introduced so that not a single fucking bug gets through. Besides, the energy of employee engagement, otherwise essential for a company’s success, is now directed at some utter bullshit. 🙆‍

KPIs are the breeding ground for the cobra effect. People become incredibly creative to get the good KPI. Unfortunately, this creativity often makes everything go to hell. You probably know how KPIs were the way to plan the economy in the USSR. A factory making spikes would be evaluated by how many tons they produce—and so the factory would make only large and heavy spikes. The state would change the KPI to the number of nails delivered—and the factory would make loads of very tiny nails. All the targets were met alright, though the end consumer suffered trying to get regular spikes via some spiv.

You would say, yeah, but the USSR was just not good at setting the right strategic objective. But what I’m trying to convey is that smart objectives have nothing to do with KPIs. So here’s the bottom line:

A few measurable targets can rarely adequately describe an entire system

Add to this the regular disconnection between teams and multiply the result by the struggle for influence in a company. Each crew has its own “meaningful KPIs” and rarely discusses them with others. In the issue, loads of local optimizations will occur to improve the marketing KPI, the sales KPI, and so on. Within a year, the whole mess will blow up the brains of an “effective manager” who has implemented this all. The manager will start searching for a new job (unless they’ve covered their ass with an even less effective mid-level manager).

So what’s the lesson to be learned here? We need to understand that there is no unambiguous connection between goals and KPIs and the desired state of the system. Reducing the churn won’t necessarily make the product better. A turnover increase does not always make a company stabler and richer. Getting rid of bugs gives no certainty whatsoever in the quality of a software product.

Instead of admiring attractive KPIs and shiny targets, we must look at the whole system, identify all connections, deeply analyze them, and build a sustainable, elaborate model of an organization. We must consider the effect of time on the entire system. We must dig deeper and never settle for simplifying our lives to just a few metrics. It isn’t about finding the magical, right KPI and reducing everything to it. It’s about seeing what you do as a complex system.

Alrighty! We’re smart, we know we shouldn’t turn metrics into goals. But can we avoid it? No way. We’re all ruled by Goodhart’s law, so as soon as you add the word “Key” to the metric, this metric immediately stops working.

So what can we do? I’ve got a very simple solution. All metrics should become KPIs! When all the metrics are KPIs, none of them actually is. 😎