# Before and After

This chapter started with the analysis of factors known about customers when they start, using stratification and hazard ratios. The previous section explained how to analyze factors that occur at the end of the customer relationship, by using competing risks. The final topic in this chapter is about events that happen during customers’ life cycles, technically known as *time-dependent covariates*. In particular, this section talks about survival measures before and after an event occurs during the customer lifetime.

Understanding time-dependent covariates starts to push the limits of what can be accomplished using SQL and Excel; statistical methods such as Cox proportional hazard regression continue the analysis beyond what we can accomplish with these tools. However, some interesting results are still possible.

This section discusses three techniques for understanding these types of factors. The first is to compare forecasts. The second is a brute-force approach...