A stochastic model is developed which can be used to set warranty periods so that a small percentage of items sold in a store are returned for rebates, repairs or replacements. If return to the store provides benefits to the retailer such as purchase of upgrades, ability to sell additional related goods or to sell snacks and other unrelated merchandise, while the customer is waiting, then profits may be maximized by setting warranty periods so that a larger percentage of the population will revisit the store. Results for products with initially high failure rates (TV sets, complex electronic components and cellular phones) versus products with gradually increasing failure rates (tires, batteries) are compared using simulations. Initially high rates of failure in conjunction with revenues from additional purchases that are greater than zero and initial losses that are low may be more lucrative than revenues from gradually increasing wear out replacements. Minimal levels of additional purchases, initial loss in value and rate reduction per number of periods can be established based on realistic simulation data in order for retailers and manufacturers to facilitate profit maximization using warranty policy.