BUSINESS MODELING: CAREER BRIEFING
Do you care about what drives your business outcomes? Do you want to understand the impact of your actions? How about predicting what might happen if you make an intervention? Or just predicting the future? If you do, then you will need a quantitative business model.
A “model” is really just an equation that uses observed regularities and patterns in your business data to establish what variables affect which and how.
The basic regression models described up to this point are sufficient for most business applications. There are some occasions though when the patterns in your data will require some extended techniques. We cover a couple of the most important of these in this chapter.
Over the past 20 years, there have been big developments in modeling business processes. Regression is an old technique but is very widely applicable and is now easily available in desktop packages, like StatproGo. If you need something more sophisticated, there are plenty of new models available, but you will probably want to consult an expert.
The regression techniques we have used so far are pretty flexible in accounting for most of the ways business inputs can affect the business output. We can handle curvilinear effects, delayed effects, binary features, and segmenting variables. However, there are still some common patterns that we cannot handle with the methods ...