
‘…And this is the group that no one ever comes into Career Day and addresses, AND IT’S CRIMINAL!’ – The Great Philosopher, Randolph Dupree
Questions and ideas are very important when working with data. Moving numbers around and just focusing on when a KPI goes up or down can be tricky. Jumping straight into analysis without asking great questions is criminal!
I want to cover two quick hypothetical scenarios and explain to you why it’s important to understand the business needs/goals and where living in a bubble of data tables can go wrong.
Would You Look at That Conversion Rate?
In this first scenario, we’re working for a dental office. One of our KPI’s is ‘conv_rate’ where we use the ‘qty_conv’ column as our numerator and ‘qty_exam’ as our denominator.
Essentially, each customer that comes in receives some sort of exam. We are wanting to track the % of customers that then move forward with some sort of dental work besides the standard exam/cleaning.
We are looking at results by month. We normally see a conversion rate in the 40–45% range.

Average spend per customer is way up as well (not pictured). We’re about to be living large!

That Doesn’t Look So Great After All
Let’s look at this from a slightly different view.

This starts to tell a different story. March and April looked like they were stellar months. Something stands out in this data. It’s 2020!
There was a lot going on at that time. As a precaution, the dentist was only seeing patients that were certain they needed dental work. Customers that were expected to be ‘exam/clean only’ were asked to book their exams for a later date.
Although our KPI looked to be improving at an astronomical rate, the end result was that we were basically only chasing those ‘slam dunk’ customers. For future business, this strategy might not work best. Let’s just hope we can cancel that customized yacht contract before it’s too late!
Something’s Driving Sales Like Crazy!
In this scenario, we’re working for a home generator company in Australia. We’re asked to look at some results of an e-mail campaign. We start to see a big uptick in sales that look like they’re coming in just after these e-mails were sent out. We should probably send truck loads of cash to this new third party marketing company for even more future growth.
‘Great success! ‘— Borat

How much influence did the e-mail campaign have? What sort of event are people thinking might happen that would drive them to buy more generators? Why are shipping costs going up like crazy as well?
A huge chunk of these orders are shipping to the US and Snovid played a huge role in all of this?!?
‘Damn the devil!’ – Diablo’s Superstar, Elliot Richards
This scenario isn’t necessarily about a tricky looking KPI. Sales going up is great. The focus is on what would drive the uptick. This might fall into the ‘correlation doesn’t equal causation’ realm.
Final Thoughts
You won’t always have the time or resources to pick the finest fruit from the top of the tree in the data analytics/science world. Just make sure that you aren’t picking up the rotting fruit from the ground and trying to pass it off as a shiny, tasty treat.
Asking questions in a group spurs more questions and more thoughts. When these questions and thoughts flow, the needle on the compass really starts to point in the right direction.
And as always, keep on learnin’!