Dynamic ranking of recommendation pairings

Abstract

A graphical user interface (GUI) may be provided by a computing system that implements a database system for presentation at a client device. The GUI may display a designated one or more criteria for selecting one of a plurality of recommendations for a target object instance associated with a designated object definition. A predictive model for determining a propensity score for selected ones of the plurality of recommendations in association with the target object instance may be configured. The propensity score may be a function of one or more data field values associated with the target object instance and may be configured based on user input received via the graphical user interface. The predictive model may be stored on a storage medium for retrieval when selecting recommendations in response to requests received to access instances of the designated object definition.

1 . A method comprising:
 transmitting via a communications interface one or more instructions for providing graphical user interface (GUI) for presentation at a client device, the graphical user interface displaying a designated one or more criteria for selecting one of a plurality of recommendations for a target object instance associated with a designated object definition;   configuring a predictive model for determining a propensity score for selected ones of the plurality of recommendations in association with the target object instance, the propensity score being a function of one or more data field values associated with the target object instance, the predictive model being configured based on user input received via the graphical user interface; and   storing the predictive model on a storage medium for retrieval when selecting recommendations in response to requests received to access instances of the designated object definition.   
     
     
         2 . The method recited in  claim 1 , wherein each propensity score indicates a respective probability that the selected one of the plurality of recommendations is accepted. 
     
     
         3 . The method recited in  claim 1 , wherein the designated one or more criteria include a designated filter rule associated with a designated one of the recommendations, the designated filter rule specifying a first criterion that must be met by the target object instance in order to select the designated one of the recommendations for presentation in association with the target object instance. 
     
     
         4 . The method recited in  claim 3 , wherein the designated one or more criteria further include a static ranking rule, and wherein the static ranking rule orders the plurality of recommendations based on a predetermined hierarchy. 
     
     
         5 . The method recited in  claim 4 , wherein the designated filter rule and the static ranking rule are stored on a storage medium in association with a recommendation strategy for retrieval when selecting recommendations in response to requests received to access instances of the designated object definition. 
     
     
         6 . The method recited in  claim 1 , wherein configuring the predictive model comprises determining at least one recommendation of the plurality of recommendations at the GUI to associate the at least one recommendation with the predictive model. 
     
     
         7 . The method recited in  claim 1 , the method further comprising:
 selecting the designated object definition at the GUI; and   selecting one or more object fields associated with the designated object definition at the GUI for use in determining the propensity score for the selected object definition.   
     
     
         8 . The method recited in  claim 7 , the method further comprising:
 segmenting the designated object definition to define a prediction group based on a designated one of the object fields.   
     
     
         9 . The method recited  claim 1 , wherein the designated object definition is a contact object associated with an account object. 
     
     
         10 . The method recited in  claim 1 , wherein configuring the predictive model comprises:
 designating a first proportion specifying a relative size of a first set of object instances as compared to a second set of object instances;   wherein when it is determined that the predictive model is unavailable,
 a respective first message is transmitted via a communication interface for each of the first set of object instances, each respective first message including a respective first one of the recommendations selected at random from a respective subset of the recommendations, and 
 a respective second message is transmitted via the communication interface for each of the second set of object instances, each respective second message including a respective second one of the recommendations, the respective second one of the recommendations determined based on a static ranking rule, the static ranking rule applying the designated one or more criteria to one or more object fields associated with the respective object instance. 
   
     
     
         11 . The method recited in  claim 10 , wherein configuring the predictive model comprises:
 designating a second proportion specifying a relative size of a third set of object instances as compared to a fourth set of object instances;   wherein when it is determined that the predictive model is available,
 a respective third message is transmitted via the communication interface for each of the third set of object instances, each respective third message including a respective first third of the recommendations selected at random from the respective subset of the recommendations, and 
 a respective fourth message is transmitted via the communication interface for each of the fourth set of object instances, each respective fourth message including a respective fourth one of the recommendations, the respective fourth one of the recommendations is selected based on a corresponding propensity score for the respective object instance. 
   
     
     
         12 . The method recited in  claim 11 , wherein each of the respective fourth one of the recommendations meets the designated one or more criteria associated with the respective object instance with which it is associated. 
     
     
         13 . The method recited in  claim 12 , wherein the predictive model is dynamically updated to include additional training data based on a plurality of responses, each response corresponding to a respective one of the recommendations. 
     
     
         14 . The method recited in  claim 13 , wherein the predictive model is available when the predictive model has been updated to include a sufficient amount of training data. 
     
     
         15 . The method recited in  claim 13 , wherein each response includes at least one object field associated with a respective one of the first set of object instances or a respective one of the second set of object instances. 
     
     
         16 . The method recited in  claim 13 , wherein each response includes at least one object field associated with the respective one of the recommendations. 
     
     
         17 . A database system implemented using a server system, the database system configurable to cause:
 transmitting via a communications interface one or more instructions for providing graphical user interface (GUI) for presentation at a client device, the graphical user interface displaying a designated one or more criteria for selecting one of a plurality of recommendations for a target object instance associated with a designated object definition;   configuring a predictive model for determining a propensity score for selected ones of the plurality of recommendations in association with the target object instance, the propensity score being a function of one or more data field values associated with the target object instance, the predictive model being configured based on user input received via the graphical user interface; and   storing the predictive model on a storage medium for retrieval when selecting recommendations in response to requests received to access instances of the designated object definition.   
     
     
         18 . The computing system recited in  claim 17 , wherein each propensity score indicates a respective probability that the selected one of the plurality of recommendations is accepted. 
     
     
         19 . The computing system recited in  claim 17 , wherein the designated one or more criteria include a designated filter rule associated with a designated one of the recommendations, the designated filter rule specifying a first criterion that must be met by the target object instance in order to select the designated one of the recommendations for presentation in association with the target object instance. 
     
     
         20 . One or more non-transitory computer readable media having instructions stored thereon for performing a method, the method comprising:
 transmitting via a communications interface one or more instructions for providing graphical user interface (GUI) for presentation at a client device, the graphical user interface displaying a designated one or more criteria for selecting one of a plurality of recommendations for a target object instance associated with a designated object definition;   configuring a predictive model for determining a propensity score for selected ones of the plurality of recommendations in association with the target object instance, the propensity score being a function of one or more data field values associated with the target object instance, the predictive model being configured based on user input received via the graphical user interface; and   storing the predictive model on a storage medium for retrieval when selecting recommendations in response to requests received to access instances of the designated object definition.