Digital advertising platform with demand path optimization
A digital advertising system includes at least one processor configured to execute a plurality of functional modules including an analytics module to receive and analyze client attributes associated with a website visitor and a requested website to define an analytics event. The analytics module ingests and enriches data within the analytics event and provides it to a machine learning module that generates prediction models for potential bids. A management platform receives the bidding prediction and generates candidate configs. An optimization module receives the candidate configs and applies weights and additional features to select a config and generate an optimized script for the selected config. A deployment module receives the optimized script and delivers the script to the website visitor.
1 . A digital advertising system, the system comprising:
one or more processors configured to execute operations comprising:
receiving client attributes associated with a client;
defining an analytics event based on analyzing the client attributes;
generating a prediction model using a machine learning module, the prediction model indicates a bid value maximization prediction for the analytics event, the machine learning module comprising a scheduler configured for periodically generating a model data set and an updated model in advance of generating the prediction model;
generating a candidate config based on the bid value maximization prediction;
generating a script associated with the candidate config; and
delivering the script to the client.
2 . The system of claim 1 , the operations further comprising:
augmenting and formatting data within the analytics event;
generating an enriched analytics event; and
generating the prediction model based on the enriched analytics event.
3 . The system of claim 1 , the operations further comprising:
providing a user interface that includes selection screens for entering bidding preferences and ad characteristics.
4 . The system of claim 1 , wherein the model data set is a bit rate model data set and the updated model is a bit rate model.
5 . The system of claim 1 , wherein the model data set comprises a publisher's bidding preferences.
6 . The system of claim 1 , wherein the scheduler generates the updated model based on periodically accessing the model data set.
7 . The system of claim 1 , wherein the machine learning module generates prediction models comprising different configuration script options.
8 . The system of claim 1 , wherein the prediction model further comprises ad options comprising ad delivery and ad placement.
9 . The system of claim 1 , the operations further comprising:
receiving one or more candidate configs;
applying weights and additional features to select a candidate config from the one or more candidate configs, wherein the weights applied by the optimization module are determined by machine learning; and
generating a plurality of optimized scripts using the selected candidate config.
10 . The system of claim 9 , wherein the weights are applied according to a website publisher's preference, wherein the pre-determined weights are determined by site-specific thresholds.
11 . The system of claim 9 , wherein the one or more candidate configs are at least partially generated using settings entered by a human administrator via the management platform.
12 . The system of claim 9 , the operations further comprising:
selecting the selected candidate config according to a plurality of features selected from page variations, number of bidders, number of geographies, bid timeout, bidder concurrency and client device type.
13 . One or more non-transitory computer-readable media having computer-executable instructions that, when executed by a computing system having a processor and memory, cause the processor to perform operations, the operations comprising:
causing display of a user interface comprising visualizations;
receiving a setting via the user interface;
using the setting, generating a candidate config for a bid value maximization, wherein the config is associated with a prediction model that indicates a maximal bid value configuration prediction for an analytics event, the prediction model is generated using a machine learning module comprising a scheduler configured for periodically generating a model data set and an updated model in advance of generating the prediction model;
generating a script associated with the candidate config; and
delivering the script to a client associated with the analytics event.
14 . The media of claim 13 , wherein the model data set is a bit rate model data set and the updated model is a bit rate model.
15 . The media of claim 13 , wherein the scheduler generates the updated model based on periodically accessing the model data set.
16 . The media of claim 13 , wherein the machine learning module generates prediction models comprising different configuration script options.
17 . A computer-implemented method, the method comprising:
generating a candidate config for a bid configuration, wherein the bid configuration is associated with a prediction model that indicates a probably of bid value maximization for an analytics event, the prediction model is generated using a machine learning module comprising a scheduler configured for periodically generating a model data set and an updated model in advance of generating the prediction model;
generating a script associated with the candidate config; and
delivering the script to a client associated with the analytics event.
18 . The method of claim 17 , wherein the model data set is a bit rate model data set and the updated model is a bit rate model.
19 . The method of claim 17 , wherein the scheduler generates the updated model based on periodically accessing the model data sets.
20 . The method of claim 17 , wherein the machine learning module generates prediction models comprising different configuration script options.