Method and apparatus for advertising bidding

1. A method for determining a bidding strategy for placing a plurality of bids for a plurality of marketing options, the method comprising:

performing, by one or more computing devices:

determining empirical data associated with marketing options based on performance metrics of the marketing options and based on observation of web site traffic from user devices to computing resources that expose the user devices to the marketing options;

generating a predictive model that comprises one or more statistical models, wherein the predictive model is generated based on the empirical data;

determining at least one modeling parameter for the predictive model, wherein the at least one modeling parameter is variable and is associated with at least one of a user characteristic or a marketing option characteristic;

determining at least one objective for the predictive model to optimize;

receiving a trigger event to optimize the predictive model, the trigger event associated with web site traffic to a set of web sites;

optimizing the predictive model by solving an objective function based in part on the at least one modeling parameter, the at least one objective, and at least one constraint; and

determining, in real time relative to the trigger event, the bidding strategy based on the optimization of the predictive model, wherein the bidding strategy comprises a set of the marketing options that results in the optimization in accordance with the at least one modeling parameter, the at least one objective, and the at least one constraint, the set of the marketing options used in connection with subsequent web site traffic to the set of web sites.

2. The method of claim 1 , further comprising placing the plurality of bids for the plurality of marketing options in accordance with the bidding strategy.

3. The method of claim 2 , wherein determining at least one objective for the predictive model to optimize includes determining a plurality of objectives for a plurality of web sites or auctions, and wherein placing the plurality of bids for the plurality of marketing options in accordance with the bidding strategy includes placing bids on different web sites.

4. The method of claim 1 , wherein the at least one objective to optimize is one of profits, conversions, revenue and traffic.

5. The method of claim 1 , wherein performing, by the one or more computing devices, further comprises:

determining a constraint for the predictive model, wherein the constraint is selected from the group consisting of money spent on marketing, cost per click, and cost per conversion; and

wherein optimizing the predictive model based in part on the at least one modeling parameter and the at least one objective further includes configuring the predictive model based upon the at least one constraint.

6. The method of claim 1 , wherein the at least one modeling parameter is one of cost of acquisition, interests and likes, geographic location, and behavioral information.

7. The method of claim 1 , wherein determining at least one objective for the predictive model to optimize includes determining objectives for multiple web sites or auctions and wherein placing the plurality of bids for the plurality of marketing options in accordance with the determined bidding strategy includes placing bids on different web sites.

8. The method of claim 1 , further comprising:

configuring a graphical user interface to receive updated model parameters;

performing a second optimization of the at least one objective in accordance with the updated model parameters;

using results of the second optimization to determine a second bidding strategy.

9. The method of claim 1 , further comprising outputting results of the optimization as a visualization.

10. A non-transitory computer-readable storage medium storing program instructions executable on a computer to implement a bidding strategy system configured for:

determining empirical data associated with marketing options based on performance metrics of the marketing options and based on observation of web site traffic from user devices to computing resources that expose the user devices to the marketing options;

generating a predictive model that comprises one or more statistical models, wherein the predictive model is generated based on the empirical data;

determining at least one modeling parameter for the predictive model, wherein the at least one modeling parameter is variable and is associated with at least one of a user characteristic or a marketing option characteristic;

determining at least one objective for the predictive model to optimize;

receiving a trigger event to optimize the predictive model, the trigger event associated with web site traffic to a set of web sites;

optimizing the predictive model by solving an objective function based in part on the at least one modeling parameter, the at least one objective, and at least one constraint; and

determining, in real time relative to the trigger event, a bidding strategy based on the optimization of the predictive model, wherein the bidding strategy comprises a set of the marketing options that results in the optimization in accordance with the at least one modeling parameter, the at least one objective, and the at least one constraint, the set of the marketing options used in connection with subsequent web site traffic to the set of web sites.

11. The non-transitory computer-readable storage medium of claim 10 , wherein the at least one objective to optimize is one of profits, conversions, revenue and traffic.

12. The non-transitory computer-readable storage medium of claim 10 further comprising determining a constraint for the predictive model wherein the constraint is one of money spent on marketing, cost per click and cost per conversion and wherein optimizing the predictive model based in part on the at least one modeling parameter and the at least one objective further includes configuring the predictive model based upon the constraint.

13. The non-transitory computer-readable storage medium of claim 10 , wherein the at least one modeling parameter is one selected from cost of acquisition, interests and likes, geographic location and behavioral information.

14. The non-transitory computer-readable storage medium of claim 10 , further comprising placing a plurality of bids for a plurality of marketing options in accordance with the bidding strategy including placing bids on different web sites and wherein the predictive model specifies model objectives for multiple web sites or auctions.

15. The non-transitory computer-readable storage medium of claim 10 , wherein performing the optimization includes performing the optimization in real time to determine the bidding strategy.

16. The non-transitory computer-readable storage medium of claim 10 , further comprising outputting results of the optimization as a visualization.

17. A system, comprising:

a non-transitory computer-readable storage medium; and

one or more processors coupled to the non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores program instructions executable by the one or more processors to implement a bidding strategy system comprising:

a modeling component for:

determining empirical data associated with marketing options based on performance metrics of the marketing options and based on observation of web site traffic from user devices to computing resources that expose the user devices to the marketing options;

generating a predictive model that comprises one or more statistical models, wherein the predictive model is generated based on the empirical data;

determining at least one modeling parameter for the predictive model, wherein the at least one modeling parameter is variable and is associated with at least one of a user characteristic or a marketing option characteristic; and

an optimizer component for:

determining at least one objective for the predictive model to optimize;

receiving a trigger event to optimize the predictive model, the trigger event associated with web site traffic to a set of web sites;

optimizing the predictive model by solving an objective function based in part on the at least one modeling parameter, the at least one objective, and at least one constraint; and

determining, in real time relative to the trigger event, a bidding strategy based on the optimization of the predictive model, wherein the bidding strategy comprises a set of the marketing options that results in the optimization in accordance with the at least one modeling parameter, the at least one objective, and the at least one constraint, the set of the marketing options used in connection with subsequent web site traffic to the set of web sites.

18. The system of claim 17 , wherein the optimizer component is further for determining a constraint for the predictive model wherein the constraint is one selected from money spent on marketing, cost per click and cost per conversion; and wherein the optimizing the predictive model based in part on the at least one modeling parameter and the at least one objective further includes configuring the predictive model based upon the constraint.

19. The system of claim 17 , wherein the predictive model specifies model objectives for multiple web sites or auctions and wherein placing a plurality of bids for a plurality of marketing options in accordance with the bidding strategy includes placing bids on different web sites.

20. The method of claim 1 , wherein the trigger event comprises at least one of: a marketer request for the optimization or a user device access to a computing resource that exposes the user device to the marketing options.