Three Key Aspects Of A Good KPI (2024)

Everything was hunky-dory in Robin’s team. The engineers were always performing at their peak and sprints were regularly completed by their deadlines. The team’s KPIs were above 80%, and everyone looked happy. All seemed perfect, maybe too good to be true?

It wasn’t until a couple of quarters down the line that the proverbial mess hit the fan. Nobody saw it coming, especially not Robin. The truth was that the team had been rushing to complete those deadlines, taking shortcuts along the way. In the process, they accumulated a lot of code debt, and that eventually came back to bite them in the rear. The product was more and more unstable, and the team had to slow down considerably to ground themselves, rework the code and catch up with the features they were supposed to implement.

What went wrong?

Key Performance Indicators (KPI) are one of the staples of modern, result-driven management. However, just implementing some random measurements doesn’t guarantee that the team will perform well. Management 101 says you need to know what to measure, how to measure it, and how to interpret the results. That’s not enough. Even if your KPI meets those three requirements, you can still end up with metrics that aren’t effective. That then leads to cases like the one above: your numbers tell a story which doesn’t match the team’s reality. In fact, it might not even be a story that your team thinks they can drive. If that’s the case, then what’s the point in having them?

Today, I want to look at how we can set up good KPIs, metrics that drive both team and business growth. These KPIs always exhibit three key aspects: relevance, measurability and simplicity.

Let’s see what that means in practice.

The “Key” in Key Performance Indicator is a solid hint to the first important aspect of a good metric: relevance. A relevant KPI tracks a core aspect of the team’s operations that can be directly matched to a company objective. Both components are critical. You can’t have a KPI that ignores a company objective, nor one that is in contradiction to the operational standards of the team.

To start, avoid measuring stuff that doesn’t matter just because a textbook tells you to, and instead measure those things that impact the bottom line. Let’s look at an example from software engineering to clarify this statement.

Say you’re making a web-based software as a service (SaaS) product, and your customer acquisition / retention is directly correlated to the first time experience they have when accessing the service. If your web app takes too long to startup, you most likely lose a customer.

A good KPI for the team is to keep this “boot up from cold” time low, say less than 20 seconds over a 1Mbps connection regardless of server load. You could argue this is a “product KPI” rather than a “team KPI”, but in the end this labeling doesn’t matter. A metric like this indirectly tracks code efficiency, which is core to the software engineering team’s operation, while providing a business benefit. It’s much better than tying code efficiency to artificial performance metrics such as CPU time / usage on the cloud provider your code runs on. You’d be optimizing for operational cost, which might not be as important as acquiring or retaining the customer.

Your KPI needs to be objectively measurable. In other words, given a set of rules, anybody measuring the team against this KPI should come up with the same final score. And while you’re at it, also make that score a single, finite number, and provide an indication of its meaning.

For instance, let’s take something that it’s already easy to measure with pure numbers: the revenue brought in by your sales representatives over a period of time, say one month. That number in itself isn’t a good KPI as there’s no scale to measure it against.

To make it finite, we can look at it as a percentage of the total revenue for the month, but that would pit your sales reps against one another. That’s not a good idea. Instead, you can look at each rep’s performance against their own results for the previous month, and provide a key of interpretation based on the overall projection for sales growth. For example, you could say that a rep has met their KPI if their sales have grown by 5% or more, while they had an excellent month if they reached 8% or more. This is a good KPI, the sales figure alone isn’t.

“Measurability is more than just numbers”

In domains where emotions are involved such as customer satisfaction, KPIs might be trickier to measure. However, there are systems that can help to translate this inherently subjective data into objective measures. For instance, you can adopt measurement scales such as the net promoter score (NPS), or you can rely on AI systems which perform sentiment analysis on your customer’s messages (find out more here on how that could work).

Whichever way you go, it’s important that the number you come up with is relatable. The person or group being measured needs to understand how they can directly influence their score. In the sales case above, the monthly growth is easy to calculate and the goal is self-explanatory: 5% or more. For the customer satisfaction case, assuming you choose to use NPS, the team needs to understand how this works and how their behavior can directly impact the final score.

In the end, KPIs are not there just to tell us how the business is going, they’re a tool to drive better performance by fostering the change that leads to it.

This is a continuation of the relatability concept above. A good KPI is simple, so that it’s easy to understand and to digest by both the producer (the team / person being measured) and the consumer (the stakeholder reading it).

On the producer side, this simplicity is what drives the strive for better performance. When the team / person being measured understands how the KPI works, they can think of ways to meet and exceed it. On the other hand, if the KPI is obscure and cannot be directly tied to day-to-day work, it’s going to end up being taken passively, or even be ignored altogether.

As an example, I’ve seen this case before, albeit a good number of years ago, in one of the engineering teams I worked in. Our performance bonuses were tied exclusively to sales performance, and as a result we felt we had no direct control over that metric. Expectedly, most of us engineers largely ignored that metric, and instead focused on things we could control, such as performance and stability of the software. There was no drive at the grassroot level to make the software better for the customer, and although sales in the end turned out to be OK — we survived — this passive approach was a lost opportunity to make revenue even better, and probably even get a higher bonus.

On the consumer side, simplicity needs to manifest itself in the relationship between the KPI and the overarching business goals it drives. The stakeholder reading your KPI report needs to understand what that means in the context of the bigger picture. They’re less interested in the raw metric than they are into what it actually means. This is why the key of interpretation we briefly mentioned in the previous section is important. However, this key alone isn’t enough. It also needs to be straightforward for the busy stakeholder to understand.

Of course, presentation is also helpful here. You can “dress up” your KPI to make it easier to digest, for instance using different colors in your report to represent lacking, acceptable or excellent performance.

KPIs can do much more for you if they’re fine tuned to your specific environment. As your company objectives evolve, so should your KPIs. As your team matures, so should your KPIs. Review your KPIs regularly to ensure they are still business-relevant.

A good KPI is never created in a vacuum. It needs the buy-in of the producer, the person or group being measured, as much as that of the consumer, the stakeholder you’re producing the measure for. Don’t be afraid to ask questions such as “does this work for you” or “how can we measure this better” to the producer, as they ultimately know their domain better than anyone else.

Also, for KPIs less is more. It’s better to have fewer KPIs which are more relevant to your business goals, than more KPIs that measure every single nook and cranny of your team. The key objective is to focus the team’s attention to what matters the most. This is not to say that aspects not covered by KPIs can be ignored. It just means that they aren’t the primary focus of your team’s growth right now. By all means, you can collect as many metrics as you want, but remember to only employ those business-critical ones as your team’s KPIs.