Methods and systems for dynamic auction floors

1. A method comprising:

tracking, by an advertising platform of a processing system including a processor, over a plurality of historical, real-time bidding auctions, bids to serve advertisements to different impression consumers, the advertising platform comprising server hardware configured to execute the real-time bidding auctions for impression inventory;

calculating over time, by the advertising platform, for each different impression consumer, a respective value of that impression consumer based on the bids to serve advertisements to that impression consumer;

storing, by the advertising platform, values of the different impression consumers in a data store accessible to the advertising platform in real-time during the real-time bidding auctions;

receiving, by the advertising platform, a request, at the advertising platform, for an advertisement to be served to an impression consumer of the different impression consumers;

in response to the request, executing, by the advertising platform, a real-time bidding auction to serve the advertisement to the impression consumer;

during the real-time bidding auction, by the advertising platform:

retrieving the respective value of the impression consumer stored in the data store;

adjusting the respective value of the impression consumer according to an impression consumer bias to obtain a biased respective value;

modifying the biased respective value according to an attribute of the impression consumer to obtain a modified, biased respective value; and

calculating a maximum floor price for the real-time bidding auction based at least in part on the modified, biased respective value of the impression consumer and one or more attributes associated with the real-time bidding auction,

wherein calculating the maximum floor price comprises respectively applying a plurality of scalar coefficients obtained from a low-latency data store to the biased respective value of the impression consumer and the one or more attributes including applying a particular scalar coefficient of the plurality of scalar coefficients to an inventory session frequency, and

wherein the inventory session frequency comprises a number of times the impression consumer has been paired with a particular impression seller over a particular time period;

identifying an aggressiveness value;

adjusting the maximum floor price according to the aggressiveness value; and

offline from the real-time bidding auction, by the advertising platform:

determining values for the plurality of scalar coefficients that, as part of calculating the maximum floor price, result in calculating an optimal floor price that maximizes revenue; and

storing the plurality of scalar coefficients in the low-latency data store.

2. The method of claim 1 , wherein the value of the impression consumer comprises (i) a value of the impression consumer to a single impression seller or (ii) a value of the impression consumer to a plurality of impression sellers.

3. The method of claim 1 , wherein the value of the impression consumer is calculated based at least in part on one or more recent bids to serve an advertisement to the impression consumer.

4. The method of claim 1 , wherein the attribute of the impression consumer is based on one of a demographic segment to which the impression consumer belongs, a geographic location of the impression consumer, or a browsing history of the impression consumer, and wherein the one or more attributes associated with the real-time bidding auction comprise one or more of time, day, publisher, average price of the impression consumer, advertisement viewing history of the impression consumer, browsing history of the impression consumer, and inventory characteristics.

5. The method of claim 1 , wherein calculating the value of a particular impression consumer further comprises:

receiving, by the advertising platform, an indication that the particular impression consumer cannot be uniquely identified;

identifying, by the advertising platform, a geography associated with the particular impression consumer; and

determining the value of the particular impression consumer based on a value of other unidentified impression consumers associated with the geography.

6. The method of claim 1 , wherein the maximum floor price comprises a hard floor, and wherein the aggressiveness value comprises a manually set aggressiveness value.

7. The method of claim 1 , wherein calculating the value of a particular impression consumer further comprises calculating, by the advertising platform, in substantially real time a moving average based on a current user value and one or more previously determined user values.

8. The method of claim 1 , wherein the value of a particular impression consumer is calculated based in part on a highest bid price and a second-highest bid price from bids received in one or more previous real-time bidding auctions.

9. The method of claim 8 , further comprising calculating, by the advertising platform, a weighted average of the highest bid price and the second-highest bid price, and wherein the value of the particular impression consumer is determined based in part on the weighted average.

10. The method of claim 1 , wherein calculating the maximum floor price further comprises determining, by the advertising platform, a function that maximizes revenue based at least in part on the value of the impression consumer and the one or more attributes associated with the real-time bidding auction, and wherein the impression consumer bias is further based on a technical attribute of a creative.

11. The method of claim 10 , wherein the revenue is determined based on a second price auction function, and wherein the technical attribute of the creative comprises one of a file type of the creative, a size of the creative, or a mode of display of the creative.

12. The method of claim 1 , wherein the maximum floor price is calculated in real time during the real-time bidding auction.

13. A system comprising:

an advertising platform of a processing system including a processor; and

a memory that stores executable instructions that, when executed by the processing system facilitate performance of operations, the operations comprising:

tracking, by the advertising platform, over a plurality of historical, real-time bidding auctions, bids to serve advertisements to different impression consumers;

calculating over time, by the advertising platform, for each different impression consumer, a respective value of the impression consumer based on the bids to serve advertisements to that impression consumer;

storing values of the different impression consumers in a data store accessible to the advertising platform in real-time during the real-time bidding auctions;

receiving a request, at the advertising platform, for an advertisement to be served to an impression consumer of the different impression consumers;

in response to the request, executing, by the advertising platform, a real-time bidding auction to serve the advertisement to the impression consumer;

during the real-time bidding auction, by the advertising platform:

retrieving a calculated value of the impression consumer stored in the data store;

adjusting the respective value of the impression consumer according to an impression consumer bias to obtain a biased respective value of the impression consumer;

modifying the biased respective value according to an attribute of the impression consumer to obtain a modified, biased respective value; and

calculating a maximum floor price for the real-time bidding auction based at least in part on the modified, biased respective value of the impression consumer and one or more attributes associated with the real-time bidding auction,

wherein calculating the maximum floor price comprises respectively applying a plurality of scalar coefficients obtained from a low-latency data store to the biased respective value of the impression consumer and the one or more attributes including applying a particular scalar coefficient of the plurality of scalar coefficients to an inventory session frequency, and

wherein the inventory session frequency comprises a number of times the impression consumer has been paired with a particular impression seller over a particular time period; and

offline from the real-time bidding auction, by the advertising platform;

identifying an aggressiveness value;

adjusting the maximum floor price according to the aggressiveness value;

determining values for the plurality of scalar coefficients that, as part of calculating the maximum floor price, result in calculating an optimal floor price that maximizes revenue; and

storing the plurality of scalar coefficients in the low-latency data store.

14. The system of claim 13 , wherein the value of the impression consumer comprises (i) a value of the impression consumer to a single impression seller or (ii) a value of the impression consumer to a plurality of impression sellers.

15. The system of claim 13 , wherein the value of the impression consumer is calculated based at least in part on one or more recent bids to serve an advertisement to the impression consumer.

16. The system of claim 13 , wherein the one or more attributes associated with the real-time bidding auction comprise one or more of time, day, publisher, average price of the impression consumer, advertisement viewing history of the impression consumer, and browsing history of the impression consumer.

17. The system of claim 13 , wherein calculating the value of a particular impression consumer further comprises:

receiving an indication that the particular impression consumer cannot be uniquely identified;

identifying a geography associated with the particular impression consumer; and

determining the value of the particular impression consumer based on a value of other unidentified impression consumers associated with the geography.

18. The system of claim 13 , wherein the maximum floor price further comprises a hard floor.

19. The system of claim 13 , wherein calculating the value of a particular impression consumer further comprises calculating in substantially real time a moving average based on a current user value and one or more previously determined user values.

20. The system of claim 13 , wherein the value of a particular impression consumer is calculated based in part on a highest bid price and a second-highest bid price from bids received in one or more previous real-time bidding auctions.

21. The system of claim 20 , wherein the operations further comprise calculating a weighted average of the highest bid price and the second-highest bid price, and wherein the value of the particular impression consumer is determined based in part on the weighted average.

22. The system of claim 13 , wherein calculating the maximum floor price further comprises determining a function that maximizes revenue based at least in part on the value of the impression consumer and the one or more attributes associated with the real-time bidding auction.

23. The system of claim 22 , wherein the revenue is determined based on a second price auction function.

24. The system of claim 13 , wherein the maximum floor price is calculated in real time during the real-time bidding auction.