I love riding my bike, feeling the wind whipping past me, the road racing underneath. But I’m also a huge techno geek, so it doesn’t come down to just feelings. I’m also on the saddle for that sweet, sweet performance data.
I can’t wait to get home to evaluate the details captured by my bike’s performance computer. And several rides ago, I began to notice a trend, repeated over and over again: I was going my fastest when I wasn’t pedaling at all! So if I wanted to go faster, I just needed to pedal less, right?
Of course not. We all know that there’s something wrong with the data there, even if it seems like there’s a relationship between my speed and pedaling power. And there was. I live in the Pacific Northwest and as my burning legs can tell you, there’s a fair share of hills around here. And coasting downhill is the best way to build up speed. My lack of pedaling wasn’t even a player here, just a byproduct.
What would have happened if I blindly went along with the data? You bet I’d be covered in scrapes and bruises. But unfortunately, when we’re off the bikes, too many companies confuse their speed with coasting too.
As companies across the globe attempt to make sense of the vast amounts of data they capture, they go with what the data suggests without analyzing what it’s really telling them. In your company, determining your competitive advantages—those key performance indicators (KPIs) that are really driving your success—shouldn’t be left to uneducated guesses and assumptions.
Don’t assume you’re speeding along when you’re only coasting. Your KPIs need to be quantified using tried and true analytical concepts. Don’t take your feet off the pedals right away. Ask yourself these questions whenever you’re looking at the data that’s driving your business decisions.
Is it a Causal Relationship?
First, take the time to determine if there is in fact a causal relationship between the variables that you’re analyzing. A correlation between two variables does not necessarily mean that a change in one variable affects the other. Time off the pedals does not equal speed. Causal relationships instead indicate that a particular event is the result of another. It is critical to determine if such a causal relationship is occurring between the metrics you’re analyzing.
And if it is causal, I implore you to utilize a regression analysis tool (or something similar) to determine the impact that the variable has on the other. It’s the best linear, unbiased estimate of the relationship between the dependent variable and one or more independent variables. Tools like this can help analyze cost behaviors or even predict future sales levels and other metrics.
What’s Your Relevant Range?
Another critical aspect for determining the relationship between variables and attempting to validate it is to establish the relevant range for which your causality is taking place. In a linear regression, this is visualized as the straight portion of the line. Data points outside of this range may not display the same correlation between variables, thus helping you identify the limits that have to be established.
In my biking data, there was a strong correlation between cadence and speed on a flat road. That was actually my relevant range. A deeper dive into my data—particularly my altitude—showed that my fastest times were when I was going downhill. But if I wanted to increase my overall speed, this actually didn’t matter much in my relevant range.
Start with those two questions and then, if you’re ready to take your training wheels off, take a deep dive into how the most successful CFOs create KPIs that help drive (pedal?) their business forward. Check out our Ebook “A Modern CFO’s Guide to the Right KPIs.”