The most important advise every retailer/business will hear is "Know your customer". Only when a business knows the needs, preferances and lifetsage of customer, can they advertise them effectively. It may include detecting early signs of lapsing behaviour or predicting the next category to cross sell or it can be just identifying purchase preferences of a customer. Solving these problems enable effective marketing and increased ROI for a business as it reduces cost and increases success rates of the marketing initiatives

Develop propensity models using neural networks leveraging temporal, customer purchase attributes to predict the propensity of customer to purchase in a segment or category or predict early signs of lapsing behaviour for a customer in a segment or category                   

These propensities can be used in targeted planning and marketing. Few examples are -