After all, there’s a narrative in a big set of numbers waiting to be read, but if we’re not aware of the plot and actors, we won’t make it past the first pages or we will feel like some key chapters are missing.
In this article, we intend to dive deeper into the metrics definition step mentioned in our article “7 steps to become data driven”, where, once you have stated your business motivation clearly for your Data project (as seen in ….) you then have to come up with the numbers to scratch that itch.
As a data analyst I have struggled with reworking the end product quite a few times to help users make the best use out of it, and my greatest takeaway was that there were some things I could’ve done better at the definition/planning phase. Some had to do with what was present in my reports, some had to do with what was missing in it.
Here’s a list of tips to keep in mind while planning metrics that can serve both businesses.
Don’t limit yourself only to what you know you have. Creating new data is as much a data project as reading the one you already have. It should even be part of product definitions as much as functionality. Hopefully, there’s always going to be a new sprint where a new fact can be registered and provide us with more knowledge. If we surrender and simply work with data as a byproduct of the operations department, it will never be thought of as an asset, and we will hit a ceiling sooner rather than later.
Research to match qualitative to quantitative
sure, happiness and love can’t be measured… But actions people perform when happy are quite measurable and almost as good. A lot of data consumers will likely want to know a good deal of qualitative information, and it is at this point that one should research on measurable behaviors and understand how they match qualitative data. Be empathic and think: what does a happy customer do? How about a frustrated one? Some areas of data can also benefit a great deal by adding a customer experience approach to take that empathy to re-discover your customers.
State the correlations clearly
When coming up with desired metrics, be sure to state and explain why X means happy and Y means frustrated. When you work with an existing app, it’s easier for every data consumer to challenge the assumption, and at the end of the day, data is only useful if it helps the business.
Don’t try to be over synthetic until you’re mature enough
A great metric at the wrong time is almost useless. If you are taking baby steps, parse your metrics as much as needed so an index is understood as a consequence of other more grounded numbers. And always bear in mind that newer, inexperienced data consumers will jump along at any point of the journey. You always want to keep a simple and clearly explained metric pyramid, with easy to grasp values at the base.
Following the last item, consider which numbers depend on others and make sense only in comparison, and treat all of them as a separate entity. Forward down the line, this will also inform how much of a timeframe you need to account into certain metrics. Does this make sense yearly? Monthly? Weekly? This is something that you will be able to find out if the context is clear at the start.
When defining metrics, following these tips will be a powerful guiding force moving forward. Not only will they provide more precise ways to estimate the effort and detail what you will need, but also guide some decisions that will come into play later in the game.
There are few things more frustrating than finishing a project and falling a few inches short because we were not aware of what story we were telling in the first place. At Baufest, we believe that data has always something to say, and we certainly know how to make a narrative out of it. Are you ready to listen?
See you next week!