Simple Ways to Use Statistics for Your Business
We all want to have super data science models that predict the future and can even prescribe what we should do next. I’m on this journey and will get there but there are nuggets of value on the way that you shouldn’t overlook. I have a few examples of simple things you can do today, quickly and easily.
Everyone has survey data. We all look at our scores and measure increases or decreases over time, mostly looking at the Top 2 box. If your job is to analyze this data, you go way deeper but most of the “users” just look at the trend to see if they’re on goal or not.
Most surveys have an “Overall…” question. This often is the main question everyone zooms in on and then there are several other questions for more detail. Have you ever checked the correlation between the “Overall…” question and the other questions?
Just seeing a little correlation can tell you if you’re asking the right question or if the responder really cares about this area.
If you find a question that has high correlation and is trending down or below target you have exposed a potential focus area. Stack ranking each question can help prioritize where you can have the most impact on the “Overall….” It’s just math. While it’s not perfect it’s a simple way to take a different angle at something we have looked at for years.
Booking, Revenue, Workload/Staffing models are great examples of Time Series data that you can use simple statistics to assist. There is a little complexity when you look at these data sets to evaluate which models are best. I’m fortunate to have a Data Scientist to help me with that.
However, this simple trend can help create a statistical baseline. Most businesses have very complex ways to come up with their forecasts or headcount planning process. I’m not suggesting you throw that out, but you can challenge those models to help get the color behind their forecasts. Again, it’s just math. In some cases this won’t work but use it where it does work.
So start here and then you can get more complex, expanding your models to become more accurate and aligning tighter to the business drivers. I still feel that this alone can be very powerful. You don’t have to wait for the super complex model to start having impact. It’s more about how you position your analysis with the stakeholders.
You don’t have to wait for that huge database or that perfect model either. The platform does not need to be too complex or expensive; these types of analysis can be implemented in any analytics platform already available in IT.
If possible, you should use a Data Scientist to help this evolve. Here are a few videos and links to models you may find useful if you are just getting started.