Methods and systems for optimizing messages to users of a social network

A computer program product comprising a non-transitory computer storage medium having instructions encoded thereon that, when executed by a processor, cause the processor to:

• determine a type of user of an online system, the type identifying a subset of characteristics of users of the online system;
• log one or more user activities associated with users of the online system associated with the type, the user activities including one or more selected from a group consisting of:
• responses by a user associated with the type to one or more messages sent to the user associated with the type, one or more times the user associated with the type is active on the social networking system, one or more activities performed by the user associated with the type, and any combination thereof;
• identify a message response prediction model for the type of user, the message response prediction using one or more selected from a group consisting of:
• the logged one or more user activities associated with users associated with the type, one or more attributes of messages sent to users associated with the type, times when the messages were sent by the social networking system to users associated with the type, and any combination thereof;
• identify one or more candidate messages, each candidate message including a link associated with a desired activity;
• determine a likelihood of the user associated with the type performing the desired activity included in each candidate message by applying the message response prediction model to each candidate message;
• select a candidate message based at least in part on the determined likelihoods;
• and
• send the selected candidate message from the social networking system to the user associated with the type at a specified time based at least in part on the determined likelihood.

A system comprising:

• at least one processor;
• and
• a memory storing instructions configured to instruct the at least one processor to:
• determine a type of user of an online system, the type identifying a subset of characteristics of users of the online system;
• log one or more user activities associated with users of the online system associated with the type, the user activities including one or more selected from a group consisting of:
• responses by a user associated with the type to one or more messages sent to the user associated with the type, one or more times the user associated with the type is active on the social networking system, one or more activities performed by the user associated with the type, and any combination thereof;
• identify a message response prediction model for the type of user, the message response prediction using one or more selected from a group consisting of:
• the logged one or more user activities associated with users associated with the type, one or more attributes of messages sent to users associated with the type, times when the messages were sent by the social networking system to users associated with the type, and any combination thereof;
• identify one or more candidate messages, each candidate message including a link associated with a desired activity;
• determine a likelihood of the user associated with the type performing the desired activity included in each candidate message by applying the message response prediction model to each candidate message;
• select a candidate message based at least in part on the determined likelihoods;
• and
• send the selected candidate message from the social networking system to the user associated with the type at a specified time based at least in part on the determined likelihood.