‘We’re very data driven!’ and ‘What do the numbers say we should do?’ are great examples of mindsets that have the possibility of getting an organization into some hot water. A business is more than numbers and KPI’s and data analysts and data scientists must always remember this.
Let’s run through a scenario that you might encounter. A fire alarm goes off. Someone has just noticed that a FR-KPI (For Real Key Performance Indicator) has not performed well for the end of the year. In fact, looking at the drastic change has some concern that the company might be in a downward spiral. Let’s take a look:
Eeeeekkk! This is our average profit per unit KPI. It’s slightly more important than the other 374 KPI’s we have: top 5-ish. To get this KPI we simply divide the $ of profit / the # of units sold.
After sifting through all sorts of SQL tables, Excel files, Jupyter Notebooks, and friendly colleague tips, you start to get an idea of what’s going on. You start to see that your product is now being offered up through a third party website and that’s costing your company profit for each item sold!
You realize that this super important KPI has the potential to go even lower! It’s very important for this KPI to go up. You rush to a decision maker and you tell them that your company should no longer offer your product through the third party website. You should stick to avenues that your company controls, which allow for more profit per unit sold. Pop the champagne bottle!
Someone came to you with a concern that a KPI was going down. They asked you to figure out why it was going down. You did just that. What’s the problem here?
This scenario is lacking context. It’s missing other KPI’s that could tell a different story. It’s absolutely needing a better understanding of the business as a whole and how it functions in the selling environment. The scope is too narrow.
When we look at a bigger picture, we note that since the start of working with the third party vendor that the quantity of items sold each month has increased significantly. This vendor has a great online presence and is actually helping lower our own advertising expenses. Yes, they require money in exchange for what they do for us, but they make things work out for the better in the end.
In fact, the bottom line has really started to grow like crazy since this new…
Continue reading: https://towardsdatascience.com/data-doesnt-decide-b2b796057e16?source=rss—-7f60cf5620c9—4