Predictive recommendation system using tiered feature data

1. A system, comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to:

generate tiered feature data associated with a set of impressions representing a tier group of promotions aggregated at different levels of granularity based on a promotion threshold level;

generate combined feature data representing at least one promotion from the tier group of promotions based at least in part on the tiered feature data and historical feature data representing the tier group of promotions; and

transmit, based in part on the combined feature data, a subset of promotions from the tier group of promotions to a client device to facilitate rendering of the subset of promotions via an electronic interface of the client device, the subset of promotions selected based in part on the combined feature data and for recommendation to a customer associated with the client device.

2. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:

generate the tiered feature data based on data collected from one or more electronic communications associated with the tier group of promotions.

3. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:

generate the combined feature data in response to a determination that the set of impressions satisfy a defined criterion.

4. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:

generate the combined feature data based at least in part on the tiered feature data, the historical feature data, and activation state data associated with a number of instances in which one or more consumers select a promotion related to the tier group of promotions.

5. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:

select the subset of promotions from the tier group of promotions based on a relevance score for respective promotions from the tier group of promotions, wherein the relevance score for the respective promotions is generated based on a predictive model derived from a set of features extracted from one or more data logs associated with the tier group of promotions.

6. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:

generate the combined feature data based on one or more of the tiered feature data, the historical feature data, or profile data associated with attributes of one or more consumers associated with one or more historical purchases related to the tier group of promotions.

7. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:

select the subset of promotions from the tier group of promotions based on a relevance score for respective promotions from the tier group of promotions.

8. The system of claim 1 , wherein the one or more storage devices store instructions that are operable, when executed by the one or more computers, to further cause the one or more computers to:

rank promotions from the tier group of promotions based on respective contextual relevance scores for the promotions.

9. A computer-implemented method, comprising:

generating tiered feature data associated with a set of impressions representing a tier group of promotions aggregated at different levels of granularity based on a promotion threshold level;

generating combined feature data representing at least one promotion from the tier group of promotions based at least in part on the tiered feature data and historical feature data representing the tier group of promotions; and

transmitting, based in part on the combined feature data, a subset of promotions from the tier group of promotions to a client device to facilitate rendering of the subset of promotions via an electronic interface of the client device, the subset of promotions selected based in part on the combined feature data and for recommendation to a customer associated with the client device.

10. The computer-implemented method of claim 9 , wherein generating the tiered feature data comprises generating the tiered feature data based on data collected from one or more electronic communications associated with the tier group of promotions.

11. The computer-implemented method of claim 9 , wherein generating the combined feature data comprises generating the combined feature data based on one or more of the tiered feature data, the historical feature data, or behavioral data associated with consumer behavior with respect to the tier group of promotions.

12. The computer-implemented method of claim 9 , further comprising:

generating the historical feature data based on one or more historical trends associated with the tier group of promotions.

13. The computer-implemented method of claim 9 , further comprising:

selecting the subset of promotions from the tier group of promotions based on a relevance score for respective promotions from the tier group of promotions.

14. The computer-implemented method of claim 9 , further comprising:

ranking promotions from the tier group of promotions based on respective contextual relevance scores for the promotions.

15. A computer program product, stored on a computer readable medium, comprising instructions that when executed by one or more computers cause the one or more computers to:

generate tiered feature data associated with a set of impressions representing a tier group of promotions aggregated at different levels of granularity based on a promotion threshold level;

generate combined feature data representing at least one promotion from the tier group of promotions based at least in part on the tiered feature data and historical feature data representing the tier group of promotions; and

transmit, based in part on the combined feature data, a subset of promotions from the tier group of promotions to a client device to facilitate rendering of the subset of promotions via an electronic interface of the client device, the subset of promotions selected based in part on the combined feature data and for recommendation to a customer associated with the client device.

16. The computer program product of claim 15 , wherein the instructions, when executed by the one or more computers, further cause the one or more computers to:

generate the tiered feature data based on data collected from a set of impressions associated with the promotion threshold level.

17. The computer program product of claim 16 , wherein the instructions, when executed by the one or more computers, further cause the one or more computers to:

generate the combined feature data in response to a determination that the set of impressions satisfy a defined criterion.

18. The computer program product of claim 15 , wherein the instructions, when executed by the one or more computers, further cause the one or more computers to:

generate the combined feature data based on one or more of the tiered feature data, the historical feature data, or activation state data associated with a number of instances in which one or more consumers select a promotion related to the tier group of promotions.

19. The computer program product of claim 15 , wherein the instructions, when executed by the one or more computers, further cause the one or more computers to:

select the subset of promotions from the tier group of promotions based on a relevance score for respective promotions from the tier group of promotions, wherein the relevance score for the respective promotions is generated based on a predictive model derived from a set of features extracted from one or more data logs associated with the tier group of promotions.

20. The computer program product of claim 15 , wherein the instructions, when executed by the one or more computers, further cause the one or more computers to:

generate the combined feature data based on one or more of the tiered feature data, the historical feature data, or profile data associated with attributes of one or more consumers associated with one or more historical purchases related to the tier group of promotions.