Abstract: A method of sending a recommendation to a customer's terminal device includes creating pairs of identifiers based on on-line activities of a group of users. Each pair contains a first and second identifier for items acquired and also acquired by a user. The unique pairings of identifiers are tallied to obtain their counts. The counts are then assigned weights so that the highest counted pairs for each first identifier have equal weight. Unique pairings that contain first identifiers in a list of identifiers derived from a customer's on-line activities are extracted with their counts. The unique second identifiers in the extracted pairings are compiled and the weighted counts are summed to form, based on the summed weighted counts, a list of recommendations. The list is then transmitted, from a computer through a communication network to the customer's terminal device, which presents the list as the recommendation to the customer.

Abstract: Notifications regarding content transactions may be received at a high rate in systems providing content to a large number of customers from a number of vendors. Each notification may contain multiple fields of information about the transaction. Business rules for determining amounts that should be invoiced for content transactions have fields of information for matching to nonfictions. As each notification is received: (1) a rule having fields of information that match fields of information contained within the notification is located; (2) an amount that should be invoiced for the transaction is determined from the located rule; (3) a line item in the determined amount is added to an open invoice for the vendor related to the content transaction that the notification is about; and (4) the updated open invoice may be saved. These steps may all be performed in real time, without copying or redistributing data in the original notification.

Abstract: A method of sending a recommendation to a customer's terminal device includes creating pairs of identifiers based on on-line activities of a group of users. Each pair contains a first and second identifier for items acquired and also acquired by a user. The unique pairings of identifiers are tallied to obtain their counts. The counts are then assigned weights so that the highest counted pairs for each first identifier have equal weight. Unique pairings that contain first identifiers in a list of identifiers derived from a customer's on-line activities are extracted with their counts. The unique second identifiers in the extracted pairings are compiled and the weighted counts are summed to form, based on the summed weighted counts, a list of recommendations. The list is then transmitted, from a computer through a communication network to the customer's terminal device, which presents the list as the recommendation to the customer.