I run regressions based on future spend, dependent upon the price bands the customer previously purchased from in the past year.
The analysis looks something like this.When prices increase, customer response changes. Customers who buy from Very Low price bands tend to spend less in the future per dollar spent the prior year. Notice the coefficients ... from $0.33 per dollar three years ago to $0.24 two years ago to $0.19 one year ago to $0.16 as of today. Basically, customers buying from Very Low price points are worth half as much as they used to be.
But wait - there's a good story here! Look at Very High price points ... the coefficients were $0.07 per dollar three years ago to $0.08 two years ago to $0.14 one year ago to $0.15 last year.
Let's pretend that customers spent $1,000,000 on Very Low price point items in the past year and $3,000,000 on Very High price point items in the past year.
- Very Low Price Points = (0.16-0.33)*$1,000,000 = ($170,000).
- Very High Price Points = (0.15-0.07)*$3,000,000 = $240,000.