Content Valuation Adjusting for Feed Placement Detriment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online systems face challenges in accurately determining the value of presenting content items to users due to real-time information not being available, which affects the types and positions of content items in user interfaces and can negatively impact user engagement.
Innovation Solution
The online system computes a value for presenting a content item based on its detrimental effect within a feed of content items by sending candidate items to a client device for ranking and determining placement, then adjusts this value based on user engagement data received from the client device to account for any inaccuracies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the online system determines content item values based on predetermined position rankings in the feed, then the valuation process is simple and fast, but the accuracy of valuation is reduced because real-time information such as device orientation and zoom settings cannot be considered
Solution Approach 1:
The system performs preliminary actions by sending candidate content items to the client device for ranking before final placement. The client device ranks content items based on real-time information (device orientation, zoom settings, feed configuration) and determines placement, then reports back to the online system. This preliminary ranking action at the client side enables accurate valuation while keeping the online system relatively simple.
Solution Approach 2:
The client device acts as an intermediary between the online system's content selection and the final content placement. It receives candidate content items, performs local ranking using real-time device information, determines actual placement in the feed, and reports this placement information back to the online system for accurate valuation calculation.
2Productivity
If the online system presents more content items in the feed, then user engagement opportunities increase, but the detrimental effect of displacing other content items to less prominent positions increases
Solution Approach 1:
The system implements feedback by having the client device report actual content item placement information back to the online system. The online system uses this feedback to calculate the detrimental effect of displacing content items and adjusts the value associated with presenting each content item accordingly, accounting for both user engagement opportunities and displacement effects.
Solution Approach 2:
The system changes parameters by dynamically adjusting the value associated with presenting content items based on multiple factors including feed position, device orientation, zoom settings, and displacement effects. This parameter adjustment allows the system to balance user engagement opportunities against the harmful displacement of other content items.
3Productivity
If the online system charges advertisers based on estimated value without considering actual placement, then the charging process is efficient, but revenue accuracy is reduced
Solution Approach 1:
The online system performs preliminary valuation by computing an estimated value for each content item based on bid amount and feed position before the client device ranks and places it. After receiving actual placement information from the client device, the system adjusts this preliminary valuation to reflect the actual detrimental effect, achieving both efficiency and accuracy.
Data Source
AI summary
An online system identifies an opportunity to present content to a viewing user and sends candidate content items eligible for presentation to the viewing user for display to a client device associated with the viewing user. The online system computes an estimated value associated with presenting a candidate content item to the viewing user based at least in part on a bid amount associated with the candidate content item and communicates the value to a content-providing user associated with the candidate content item. Based on information received from the client device describing the placement of one or more candidate content items within a feed of content items presented to the viewing user, the online system calculates a detrimental effect of presenting the candidate content item on user engagement with the online system by the viewing user and computes an adjusted value associated with its presentation based on the detrimental effect.