Automated on-line experimentation to measure users behavior to treatment for a set of content elements
First Claim
Patent Images
1. An automated computer-implemented method comprising:
- defining an experiment to gauge user reaction to various treatments for a set of content elements, conducting the experiment over a data network; and
collecting over the data network observation data relating to user behavior for each treatment wherein conducting the experiment comprises;
statistically sampling to specify a plurality of control groups, each control group comprising at least one user;
specifying a particular treatment to be delivered to the at least one user in each control group;
receiving identical requests for content from the respective at least one user in each of the control groups; and
in response to the identical requests, delivering to the at least one user in each control group a different treatment.
5 Assignments
0 Petitions
Accused Products
Abstract
An automated system for experimentation includes an experiment engine which can define an experiment relating to various treatments for a set of content elements. The experiment engine conducts the experiment over a data network. An observation module collects observation data relating to user behavior for each treatment. A scripting/scheduling engine coordinates the operation of the experiment engine and the observation module.
254 Citations
43 Claims
-
1. An automated computer-implemented method comprising:
-
defining an experiment to gauge user reaction to various treatments for a set of content elements, conducting the experiment over a data network; and
collecting over the data network observation data relating to user behavior for each treatment wherein conducting the experiment comprises;
statistically sampling to specify a plurality of control groups, each control group comprising at least one user;
specifying a particular treatment to be delivered to the at least one user in each control group;
receiving identical requests for content from the respective at least one user in each of the control groups; and
in response to the identical requests, delivering to the at least one user in each control group a different treatment. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10)
-
-
11. An automated system comprising:
-
a content system operable to store content, the content comprising a set of content elements; and
a communication management system in communication with the content system and operable to define an experiment relating to various treatments for the set of content elements, conduct the experiment over a data network, and collect over the data network observation data relating to user behavior for each treatment, wherein the system is configured for statistically sampling to specify a plurality of control groups, each control group comprising at least one user;
specifying a particular treatment to be delivered to the at least one user in each control group;
receiving identical requests for content from the respective at least one user in each of the control groups; and
in response to the identical requests, delivering to the at least one user in each control group a different treatment. - View Dependent Claims (12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25)
-
-
26. A computer-implemented method for on-line experimentation comprising:
-
defining a plurality of treatments for a set of content elements;
receiving requests for content from users over a data network;
statistically sampling to create at least one control group of users;
allocating over the data network a first treatment to each user in the control group;
allocating over the data network a second treatment to each user not in the control group; and
collecting observation data for observed behavior of users in the control group and users not in the control group. - View Dependent Claims (27, 28, 29)
-
-
30. An online experimentation system configured to automatically create various treatments for a set of content elements according to systematic rules of experimentation that determine a number of alternative treatments using various combinations of the content elements from the set, and to allocate the treatments to users to test user behavior in response to the various treatments, the system being further configured to capture data relating to the observed behavior of users relative to the various treatments and to perform one or more methods comprising:
- (1) analyze the data to derive a model in form of a contingency table;
(2) process the data for pathologies, including at least one of missing data, and structural dependencies, and processing the data to be model ready;
(3) make inferences about some variables that influence user behavior based on random utility theory; and
(4) modeling user behavior based on Bayesian Markov Chain Monte Carlo estimation procedure.
- (1) analyze the data to derive a model in form of a contingency table;
-
31. A machine readable medium including a set of instructions for:
-
defining an experiment to gauge user reaction to various treatments for a set of content elements;
conducting the experiment over a data network; and
collecting over the data network observation data relating to user behavior for each treatment wherein conducting the experiment comprises;
statistically sampling to specify a plurality of control groups, each control group comprising at least one user;
specifying a particular treatment to be delivered to the at least one user in each control group;
receiving identical requests for content from the respective at least one user in each of the control groups; and
in response to the identical requests, delivering to the at least one user in each control group a different treatment.
-
-
32. A content page comprising a set of content elements generated by a computer implemented system for:
-
defining an experiment to gauge user reaction to various treatments for a set of content elements;
conducting the experiment over a data network; and
collecting over the data network observation data relating to user behavior for each treatment wherein conducting the experiment comprises;
statistically sampling to specify a plurality of control groups, each control group comprising at least one user;
specifying a particular treatment to be delivered to the at least one user in each control group;
receiving identical requests for content from the respective at least one user in each of the control groups; and
in response to the identical requests, delivering to the at least one user in each control group a different treatment.
-
-
33. An automated system for experimentation, the system being configured to:
-
generate and allocate a set of treatments from a predefined set of control elements, according to experiment designs, the system capable of determining a design by selecting from design criteria comprising;
the specific effects allowed;
the number of factors and factor levels included; and
the amount of information produced in relation to the objective outcome; and
collect observation data of user behavior to the set of treatments and determine the effect of each control element on user behavior according to choice models, behavioral models, or contingency tables. - View Dependent Claims (34, 35, 36)
-
-
37. A computer implemented method comprising:
-
automatically creating various treatments for a set of content elements according to systematic rules of experimentation that determine a number of alternative treatments using various combinations of the content elements from the set;
allocating the treatments to users to test user behavior in response to the various treatments;
capturing data relating to the observed behavior of users relative to the various treatments;
performing one or more of;
(1) analyzing the data to derive a model in form of a contingency table;
(2) processing the data for pathologies, including at least one of missing data, and structural dependencies, and processing the data to be model ready;
(3) making inferences about some variables that influence user behavior based on random utility theory; and
(4) modeling user behavior based on Bayesian Markov Chain Monte Carlo estimation procedure.
-
-
38. A computer implemented method, comprising:
-
generating and allocating a set of treatments from a predefined set of content elements, according to experiment designs, the system capable of determining a design by selecting from design criteria comprising;
the specific effects allowed;
the number of factors and factor levels included; and
the amount of information produced in relation to the objective outcome; and
collecting observation data of user behavior for the set of treatments and determining the effect of each content element on user behavior according to choice models, behavioral models, or contingency tables. - View Dependent Claims (39)
-
-
40. A machine readable medium storing a set of instructions for performing the steps comprising:
-
automatically creating various treatments for a set of content elements according to systematic rules of experimentation that determine a number of alternative treatments using various combinations of the content elements from the set;
allocating the treatments to users to test user behavior in response to the various treatments capturing data relating to the observed behavior of users relative to the various treatments;
performing one or more of;
(1) analyzing the data to derive a model in form of a contingency table;
(2) processing the data for pathologies, including at least one of missing data, and structural dependencies, and processing the data to be model ready;
(3) making inferences about some variables that influence user behavior based on random utility theory; and
(4) modeling user behavior based on Bayesian Markov Chain Monte Carlo estimation procedure.
-
-
41. A machine readable medium storing a set of instructions for performing the steps comprising:
-
generating and allocating a set of treatments from a predefined set of content elements, according to experiment designs, the system capable of determining a design by selecting from design criteria comprising;
the specific effects allowed;
the number of factors and factor levels included; and
the amount of information produced in relation to the objective outcome; and
collecting observation data of user behavior for the set of treatments and determining the effect of each content element on user behavior according to choice models, behavioral models, or contingency tables. - View Dependent Claims (42)
-
-
43. A computer implemented method, comprising:
-
defining an experiment to gauge user reaction to various treatments for a set of content elements;
conducting the experiment over a data network; and
collecting over the data network observation data relating to user behavior for each treatment wherein conducting the experiment comprises;
statistically sampling to specify a plurality of control groups, each control group comprising at least one user;
specifying a particular treatment to be delivered to the at least one user in each control group;
receiving identical requests for content from the respective at least one user in each of the control groups; and
in response to the identical requests, delivering to the at least one user in each control group a different treatment.
-
Specification