Online serving threshold and delivery policy adjustment
The present invention provides techniques for use in association with online advertising, relating to use of serving thresholds, associated with predicted click through rates, and delivery policies, associated with advertising inventory serving and distribution. An offline-trained machine learning-based model may be utilized in advertising serving decision-making in connection with serving opportunities. However, serving thresholds and delivery policies, for use in association with the model in serving decision-making, may be adjusted online, such as in real-time or near real-time, based on information obtained online affecting factors such as predicted click through rates and advertising inventory distribution.
1. A method comprising:
during an offline period, initially determining a set of serving thresholds to be utilized in online advertisement serving, wherein a serving threshold is associated with a minimum anticipated click through rate;
during an offline period, initially determining a set of delivery policies, wherein a delivery policy is associated with one or more rules relating to serving of advertisements in accordance with required or optimal distribution of advertising inventory across serving opportunities;
during an online period, receiving and storing by a server coupled to the Internet, category-specific advertising information in one or more data sources;
during an online period, querying the one or more data sources to determine adjustments based on the stored category-specific advertising information;
during an online period, and based at least in part on information obtained during an online period including the stored category-specific advertising information sent from the one or more data sources over the Internet, dynamically adjusting in real-time, based on the determined adjustments, at least one of the set of serving thresholds to determine at least one adjusted serving threshold, and dynamically adjusting in real-time at least one of the set of delivery policies to determine at least one adjusted delivery policy;
during an online period, utilizing a machine learning-based model in dynamically updating decision-making in real-time with regard to serving of online advertisements in connection with serving opportunities based at least in part on the at least one adjusted serving threshold and the at least one adjusted delivery policy, adjusted dynamically during the online period, wherein the machine learning-based model is initially trained during an offline period with advertising-related information stored in and queried from the one or more data sources and dynamically updated during an online period with real-time adjustments, wherein an online period is a period of active advertisement serving in which the model is utilized; and
during an online period, transmitting an online advertisement over the Internet based on a particular serving opportunity to a user device of a particular user based on the decision-making by the machine learning-based model.
2. The method of claim 1 , comprising utilizing a hierarchical taxonomy of behavioral targeting categories in behavioral targeting of users, wherein nodes represent behavioral targeting categories, and wherein delivery policies are utilized in optimally or more optimally distributing serving of advertisement inventory across nodes of the taxonomy.
3. The method of claim 1 , comprising training the model using information including historical user behavior information in association with online advertising.
4. The method of claim 1 , wherein the at least one adjusted serving threshold and the at least one adjusted delivery policy are adjusted based at least in part on real-time or near real-time monitored click through rate information and advertising inventory distribution.
5. The method of claim 1 , wherein the at least one adjusted serving threshold and the at least one adjusted delivery policy are adjusted based at least in part on real-time or near real-time monitored click through rate information and advertising inventory distribution in association with particular nodes of a hierarchical taxonomy of behavioral targeting categories, wherein the nodes represent behavioral targeting categories.
6. The method of claim 1 , comprising utilizing serving thresholds in achieving desired levels of advertisement performance in connection with particular categories of a hierarchical taxonomy of user interest categories.
7. The method of claim 1 , comprising adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a news event occurring during an online period, wherein the news event affects user interests pertinent in behavioral targeting.
8. The method of claim 1 , comprising adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a calendar or chronologically-based circumstance occurring during an online period that effects user behavior pertinent in behavioral targeting.
9. The method of claim 1 , comprising monitoring, during an online period, category-specific advertising inventory serving distribution information and category-specific click through rate information, for use in adjusting at least serving threshold and at least one delivery policy.
10. The method of claim 1 , comprising monitoring, during an online period, category-specific advertising inventory serving distribution information and category-specific click through rate information, for use in adjusting at least serving threshold and at least one delivery policy, wherein the information is obtained from real-time or near real-time behavioral targeting and advertising serving systems.
11. The method of claim 1 , comprising adjusting at least one delivery policy to reflect an increased predicted click through rate, for a time frame and for a particular behavioral targeting category, that was predicted online during a particular online period, relative to a predicted click through rate, for the time frame and associated with the category, that was predicted during an offline period.
12. The method of claim 1 , comprising adjusting at least one delivery policy to reflect an increased predicted click through rate, for a time frame and for a particular behavioral targeting category, that was predicted online during a particular online period, relative to a predicted click through rate, for the time frame and associated with the category, that was predicted during an offline period, comprising prioritizing delivery relating to the particular behavioral targeting category during the time frame.
13. The method of claim 1 , comprising adjusting at least one delivery policy to reflect an increased predicted click through rate, for a time frame and for a particular behavioral targeting category, that was predicted online during a particular online period, relative to a predicted click through rate, for the time frame and associated with the category, that was predicted during an offline period, comprising prioritizing delivery relating to the particular behavioral targeting category during the time frame, based at least in part on a ratio of the increased predicted click through rate, predicted online, to the predicted click through rate that was predicted during an offline period.
14. The method of claim 1 , wherein serving threshold adjustment and delivery policy adjustment is performed automatically based on monitored online information.
15. The method of claim 1 , wherein serving threshold adjustment and delivery policy adjustment is performed automatically based on monitored online information, to facilitate maintaining desired click through rate and delivery distribution across behavioral categories by facilitating accounting for differences between predicted user traffic patterns predicted offline from predicted user traffic patterns predicted online based on updated real-time or near real-time information.
16. A system comprising:
one or more server computers coupled to a network and Internet; and
one or more databases coupled to the one or more server computers;
wherein the one or more server computers are configured for:
during an offline period, initially determining a set of serving thresholds to be utilized in online advertisement serving, wherein a serving threshold is associated with a minimum anticipated click through rate;
during an offline period, initially determining a set of delivery policies, wherein a delivery policy is associated with one or more rules relating to serving of advertisements in accordance with required or optimal distribution of advertising inventory across serving opportunities;
during an online period, receiving and storing by a server coupled to the Internet, category-specific advertising information in one or more data sources;
during an online period, querying the one or more data sources to determine adjustments based on the stored category-specific advertising information;
during an online period, and based at least in part on information obtained during an online period, including the stored category-specific advertising information sent from the one or more data sources over the Internet, dynamically adjusting in real-time, based on the determined adjustments, at least one of the set of serving thresholds to determine at least one adjusted serving threshold, and dynamically adjusting in real-time at least one of the set of delivery policies to determine at least one adjusted delivery policy;
during an online period, utilizing a machine learning-based model in dynamically updating decision-making in real-time with regard to serving of online advertisements in connection with serving opportunities based at least in part on the at least one adjusted serving threshold and the at least one adjusted delivery policy, adjusted dynamically during the online period, wherein the machine learning-based model is initially trained during an offline period with advertising-related information stored in and queried from the one or more data sources and dynamically updated during an online period with real-time adjustments, wherein an online period is a period of active advertisement serving in which the model is utilized; and
during an online period, transmitting an online advertisement over the Internet based on a particular serving opportunity to a user device of a particular user based on the decision-making by the machine learning-based model.
17. The system of claim 16 , wherein the one or more server computers are coupled to the Internet.
18. The system of claim 16 , comprising utilizing a hierarchical taxonomy of behavioral targeting categories in behavioral targeting of users, wherein nodes represent behavioral targeting categories, and wherein delivery policies are utilized in optimally or more optimally distributing serving of advertisement inventory across nodes of the taxonomy.
19. The system of claim 16 , comprising monitoring, during an online period, category-specific advertising inventory serving distribution information and category-specific click through rate information, for use in adjusting at least one serving threshold and at least one delivery policy.
20. A non-transitory computer readable medium or media containing instructions for executing a method for use in association with an online advertising auction relating to an online advertising marketplace, the method comprising:
during an offline period, initially determining a set of serving thresholds to be utilized in online advertisement serving, wherein a serving threshold is associated with a minimum anticipated click through rate;
during an offline period, initially determining a set of delivery policies, wherein a delivery policy is associated with one or more rules relating to serving of advertisements in accordance with required or optimal distribution of advertising inventory across serving opportunities;
wherein the distribution of advertising inventory across serving opportunities relates to distribution of advertising inventory across serving opportunities associated with different nodes of a hierarchical taxonomy of behavioral targeting categorical nodes;
during an online period, receiving and storing by a server coupled to the Internet, category-specific advertising information in one or more data sources;
during an online period, querying the one or more data sources to determine adjustments based on the stored category-specific advertising information;
during an online period, and based at least in part on information obtained during an online period, including the stored category-specific advertising information sent from the one or more data sources over the Internet, dynamically adjusting in real-time, based on the determined adjustments, at least one of the set of serving thresholds to determine at least one adjusted serving threshold, and dynamically adjusting in real-time at least one of the set of delivery policies to determine at least one adjusted delivery policy;
wherein the at least one adjusted serving threshold and the at least one adjusted delivery policy are adjusted at least in part to better optimize user targeting or advertising inventory distribution, based at least in part on the information obtained during the period;
during an online period, utilizing a machine learning-based model in dynamically updating decision-making in real-time with regard to serving of online advertisements in connection with serving opportunities based at least in part on the at least one adjusted serving threshold and the at least one adjusted delivery policy, adjusted dynamically during the online period, wherein the machine learning-based model is initially trained during an offline period with advertising-related information stored in and queried from the one or more data sources and dynamically updated during an online period with real-time adjustments, wherein an online period is a period of active advertisement serving in which the model is utilized;
wherein the machine-learning based model is trained using information including historical user behavior information; and
during an online period, transmitting an online advertisement over the Internet based on a particular serving opportunity to a user device of a particular user based on the decision-making by the machine learning-based model.