Ad Forecasting Using Historical Search Indexes for Real-Time CTR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing forecasting methods and systems for web-based advertising campaigns suffer from inaccuracy, leading to inefficient energy consumption and increased carbon footprint due to the transmission of inaccurate impressions.
Innovation Solution
A computer-implemented method and system that utilizes a forecasting service to translate targeting criteria and placement data into an index querying syntax, query historical searches, and derive a click-through-rate (CTR) in real-time, using machine learning models and large datasets to improve forecasting accuracy and reduce the number of impressions needed, thereby reducing energy consumption and bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional forecasting methods are used for web-based advertising campaigns, then the system can operate with simpler algorithms, but the forecasting accuracy deteriorates leading to inaccurate impressions and increased energy consumption
Solution Approach 1:
The system performs preliminary forecasting calculations using machine learning models before actual ad campaign execution. By predicting impressions and clicks in advance with high accuracy, the system可以避免 transmitting unnecessary ad data, thereby reducing energy consumption while maintaining or improving forecasting accuracy
Solution Approach 2:
The patent replaces traditional statistical forecasting methods with machine learning models that leverage historical data and patterns. This substitution enables more accurate predictions of ad performance metrics, reducing the need for excessive impressions and associated energy consumption
2Reliability
If more impressions are transmitted to compensate for poor forecasting accuracy, then the system can maintain conservative estimates, but the energy consumption and bandwidth usage increase
Solution Approach 1:
The system incorporates feedback loops where actual ad campaign performance data is continuously fed back into the machine learning models. This enables the models to learn from real outcomes and improve forecasting reliability over time, reducing the need for conservative over-estimation and associated energy waste
Solution Approach 2:
The patent dynamically adjusts forecasting parameters based on historical performance data and changing market conditions. By optimizing parameters such as click-through rates and impression values using machine learning, the system achieves reliable forecasts without transmitting excessive ad data
3Productivity
If real-time forecasting is implemented to improve speed, then the system can provide faster ad campaign creation, but the computational complexity and processing requirements increase
Solution Approach 1:
The system performs preliminary data processing and feature extraction during off-peak times, preparing datasets for rapid machine learning inference during real-time operations. This pre-processing reduces the computational burden during actual forecasting, enabling fast performance while managing system complexity
Solution Approach 2:
The machine learning models are designed to automatically process incoming data and generate forecasts without requiring complex manual intervention or centralized coordination. This self-service capability reduces operational complexity while maintaining real-time forecasting speed
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There is disclosed a computer-implemented method of forecasting a number of clicks and a number of impressions a line item will receive in a defined time period, the method including the steps of: (i) a forecasting service receiving a request from a portal for a forecast of the number of clicks and the number of impressions a line item will receive in a defined time period, the request including targeting criteria and placement data; (ii) the forecasting service translating the targeting criteria and the placement data into an index querying syntax; (iii) the forecasting service using the index querying syntax to request inventory via a service; (iv) the service making a request corresponding to the index querying syntax for a time series of historical searches to a database system; (v) in response to the request, the service receiving the time series of historical searches from the database system; (vi) the service providing inventory to the forecasting service corresponding to the defined time period, based on the received time series of historical searches; (vii) the forecasting service using the returned inventory in a request to a real-time service which makes a forecast of the number of clicks and the number of impressions the line item will receive in the defined time period, in which the forecasting service receives the response from the real-time service in real time, e.g. in less than one second; (viii) the forecasting service using the received forecast of the number of clicks and the number of impressions the line item will receive in the defined time period, to derive a click-through-rate (CTR) for the line item; (ix) the forecasting service returning the forecast of the number of clicks and the number of impressions the line item will receive in the defined time period, and the click-through-rate (CTR) in the defined time period, to the portal. Related methods and systems are disclosed.