The system links Internet web page context with audience usage and location data to support advertising efficiency and effectiveness. An ontology of categories is created where domains and website pages are classified and scored against the links on those pages and the meta-tag key word pools that are harvested from those web pages. An ontology of high level categories are derived from the frequency of the key words appearing within the domain URL addresses of the pages, the domain of the links on those pages or within the content of the pages themselves. A method includes building a training set of web pages from a plurality of ad networks and sites where the system captures impressions in the form of real-time bids as well as click through events that include the IP address, the domain, the time of day and day of week, ad size and position, browser type, and bid amount whereby the training set is aggregated in a database whereby successful bids can be used in combination with audience and category attributes to model and score impression bids that combine the optimal mix of audience attributes, location, categorical affinity and bid price.

An online system selects for display content items having an option to allow a user to converse with a content provider over an electronic communication system in a way that optimizes for the occurrence of that communication. Content items provided by the online system include links to an online communication system from which the online system can directly observe whether the user engaged in communications with third parties. The online system can thus obtain labeled training data describing communications between users and users' previous interactions with content items and pages of the online system. The trained model is applied to candidate content items to predict a probability that a user will engage in a communication with the content provider if the content is provided to the user, and the model optimizes the selection of content items for which the call to action is to engage in the communication.

A method, apparatus, and computer program product are disclosed to improve generation of electronic communications. The method may provide a plurality of content slots each configured to receive content, the content comprising at least one of promotion content or non-promotion content. The method may also include maintaining a database comprising a plurality of promotion content generators and non-promotion content generators, and determining, using a processor, one of the plurality of promotion content generators or non-promotion content generators for respectively supplying corresponding promotion content or non-promotion content to each of the plurality of content slots. The determining the one of the plurality of promotion content generators or non-promotion content generators may include determining selection parameters, and scoring the plurality of promotion content generators and non-promotion content generators based at least in part on the selection parameters.

An ad is placed in a movie, by analyzing inherent characteristics of the movie, analyzing viewed characteristics of the movie, analyzing viewer characteristics of a viewer of the movie, obtaining advertiser preferences for placement of the ad in the movie, determining costs of placing the ad in the movie based on the inherent characteristics of the movie, the viewed characteristics of the movie, the viewer characteristics and the advertiser preferences, and placing the ad in the movie in accordance with the inherent characteristics of the movie, the viewed characteristics of the movie, the viewer characteristics, the advertiser preferences and the determined costs.

Operations include determining a sales conversion window based on product characteristics and/or user characteristics. Product characteristics may include a type of product, a price of the product, a style of the product, and a lifetime of the product. User characteristics may include conversion windows for prior purchases by a user, browsing history, and demographic information. The system may determine the sales conversion window based on events indicative of user interest. The events may include viewing a product, liking a product, sharing a product, adding a product to a cart for purchase, and purchasing a product. One or more events may be used to generate a model for the sales conversion window. Some embodiments execute one or more actions for generating sales during the computed sales conversion window.

Disclosed embodiments relate to performing an advertisement campaign filtering process while protecting the privacy of both an advertiser and a user of a personal computing device. Techniques include maintaining a plurality of sets of advertising competition rules, the plurality of sets of advertising competition rules being associated with a plurality of discrete advertising campaigns; for a set of advertising competition rules from the plurality of sets of advertising competition rules: identifying advertisement targeting criteria associated with the set of advertising competition rules, differentiating, from within the advertisement targeting criteria, between advertisement-sensitive targeting criteria and advertiser-insensitive criteria, and transforming the advertisement-sensitive sensitive targeting criteria; and transmitting, to the personal computing device, at least a portion of the transformed advertisement-sensitive targeting criteria.

Disclosed embodiments relate to conducting a real-time selection of a targeted advertisement or campaign on a personal computing device. Techniques include receiving, on the personal computing device, a plurality of sets of advertising competition rules associated with a plurality of discrete advertising campaigns. Further techniques include identifying an opportunity to display a targeted advertisement to a user of the personal computing device, accessing sensitive data associated with the user stored locally on the personal computing device, the sensitive data not being made available outside of the personal computing device without authorization from the user, and conducting, based on the identified opportunity and using the accessed sensitive data, a real-time selection of at least one of the plurality of discrete advertising campaigns, the real-time selection is based on one or more of the plurality of sets of advertising competition rules.

Disclosed embodiments relate to efficiently providing advertising competition rules to a personal computing device of a user. Techniques include maintaining a plurality of sets of advertising competition rules, the plurality of sets of advertising competition rules being associated with a plurality of discrete advertising campaigns; receiving non-personal advertisement targeting data from the personal computing device; filtering the plurality of sets of advertising competition rules using a filtering technique, based on the non-personal advertisement targeting data, to identify a subset of the plurality of sets of advertising competition rules; and transmitting the subset of the plurality of sets of advertising competition rules to the personal computing device.

Disclosed embodiments relate to receiving a targeted advertisement on a personal computing device without revealing sensitive data of a user of the personal computing device. Techniques include maintaining sensitive data associated with the user; identifying, while the user is interacting with an application running on the personal computing device, an opportunity to display a targeted advertisement to the user of the personal computing device; receiving a prompt to request a targeted advertisement for display on the personal computing device, the targeted advertisement being part of a discrete advertising campaign selected based on the sensitive data associated with the user; and requesting, based on the prompt, the targeted advertisement for display on the personal computing device.

Methods and corresponding systems for automatic frequency capping provide an automated decision-making process that decides dynamically how often the content of electronic communications from an entity should be sent to any specific customer or potential customer for a particular campaign. An individual customer's or potential customer's optimal electronic communication frequency is determined using machine learning and is based on behavior data. The method may comprise receiving from an entity, content and an audience, that includes at least a particular customer or potential customer, for use for generating electronic communications for the particular campaign; training a model to learn a personalized frequency for sending the electronic communications to each of the audience; based on the trained model, the content, and the audience, creating electronic communications to send to each of the audience; and causing the electronic communications to be sent to each of the audience at a frequency that is personalized.