This series unpacks the different types of analytics that exist and the best applications for each. Understanding these distinct varieties better equips us to operate successfully in the analytics space and navigate its cyclical nature effectively.
Descriptive analytics is all about answering simple questions like:
- How much did we sell last week?
- Which customers bought the most from us?
- How much inventory do we have left?
It would be wrong, however, to think that answering these simple questions is itself simple. The temptation is to view this type of analytics as going to data and reporting it out. However, there are some fundamental choices we are making: what to capture, how to capture it and what level of abstraction we apply to it. These choices are important and strongly influence the overall effectiveness of our analytics.
Data-Driven Decisions
What information we capture and how we do it can actually alter our environment. When we tell our sales staff that we will track the number of deals they close, we incentivize them to close more deals. This can have all kinds of side effects, like the sales team pursuing small deals at the expense of revenue.
Decisions on data never happen in a vacuum. Data communicates what we value as a company and can alter the very things we’re seeking to describe. Take the time to discuss the best ways to describe business processes and the possible downsides of doing so. Don’t settle for a single, long-held KPI when multiple viewpoints are needed.
Grouping and Categorizing Data
Further, there’s another layer of complication since we rarely spend time on the raw data. As we do our analytics, we aggregate and summarize the data to a level that is easy to understand and communicate. This act increases our ability to understand, but it does so at the cost of losing information. If we’re looking at student grade point averages, the average hides the extremes from us. A student who averages a B might actually be an A-student hiding a D in a class they need help in.
The way we group our data and the statistics we use are all important—too high level and we miss key information; too low level and it becomes an information overload. The key to success is finding a balance, continuing to look at things from multiple angles and keeping in mind that you are looking at a simplification of the reality beneath the analysis.
The Maturity of Descriptive Analytics
Maturity within descriptive analytics (type 2) is along three dimensions:
- The speed at which you can report the data (from batches to real-time)
- The number of data sources
- The quality of that data
The next post in this series will highlight a second type of analytics that focuses on data patterns and correlations.
