The operation of an online shop that sells goods from a certain range is considered. There is a service area that can be defined by a network in which nodes of potential consumers are located. The task is to maximize profits of the participating agents or minimize the cost function of these agents. Revenue can be increased by reducing shipping costs by designing economical customer diversions routes, and by increasing the number of customers by lowering the price and placing drop-off points convenient for customers to pick up the goods themselves. This paper deals with two practical problems. The first problem involves two agents: The first is the online shop itself, which will make the delivery and wants to minimise its shipping costs. The second agent is the customer who is waiting for delivery and wants to minimise waiting time. The second problem involves three agents: The first two agents are the same as in the first problem, and a third agent appears, advertising, who needs to choose the path in which the benefits of advertising are maximised. It is convenient to solve these problems with a decision tree, whose branches will correspond to customer diversions strategies, in a given period of time, and given that in the next period there is a change in demand due to the actions of economic agents (creating advertising, detouring customers past crowded places and others), diversions strategies should also be changed.