Customers never announce they are not coming back. Machine learning identifies customer engagement patterns – purchases, opened emails, web visits, etc. Customers who depart from established patterns toward disengagement are flagged for intervention so you can save the relationship. Often, a well-timed nudge to wake them up is all that is needed.
A key, actionable pattern is the frequency of shoe purchases. Living (each month they are different) customer cohorts are identified and treated with the appropriate temperament of messaging, e.g. shoe wear quiz, repurchase reminder, or please come back.