Query Completion Intent Diversion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users often encounter query completions that divert their original search intent, leading to irrelevant results, as these completions may be amusing or entertaining but not relevant to the actual information sought.
Innovation Solution
A method is implemented to identify and demote query completions that change users' original search intent by calculating a completion origination score, testing the impact of withholding these completions, and classifying them based on a decrease in submission frequency, using a database of query completions with popularity scores to select alternative suggestions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If query completions are provided to assist users in searching, then user convenience is improved, but users may be diverted from their original search intent
Solution Approach 1:
The system performs A/B testing by withholding query completions from a test group and comparing search query submission frequencies against a control group. This feedback mechanism identifies completions that cause significant drops in relevant query submissions, indicating they divert users from original search intent. The feedback loop allows continuous refinement of completion suggestions to balance user convenience with preserving search intent.
Solution Approach 2:
The system dynamically adjusts the popularity scores of query completions based on test results. Completions classified as changing search intent have their popularity scores reduced, while those that assist users without diverting them maintain or increase their scores. This parameter change enables the system to adapt completion suggestions over time, optimizing the balance between helping users and preserving their original search intent.
2Ease of operation
If popular query completions are suggested based on frequency, then user convenience is improved, but irrelevant or amusing completions may be promoted
Solution Approach 1:
The system uses A/B testing feedback to evaluate whether popular query completions actually assist users or divert them. By measuring the impact of withholding completions on search query submission frequencies, the system identifies whether high-frequency completions are genuinely helpful or merely amusing diversions. This feedback mechanism filters out irrelevant completions even if they are popular.
Solution Approach 2:
The system dynamically changes the popularity scores of query completions based on their performance in tests. Completions that divert users from their search intent have their scores reduced, while those that genuinely assist users have their scores maintained or increased. This dynamic parameter adjustment ensures that only relevant completions are promoted, improving the reliability of suggestions.
3Reliability
If A/B testing is performed to identify intent-changing completions, then completion quality is improved, but system complexity increases
Solution Approach 1:
The system segments users into test and control groups for A/B testing, allowing independent evaluation of completion impact. This segmentation enables the system to isolate the effect of specific query completions on user behavior without affecting the entire user base. The segmented approach manages complexity by testing completions incrementally rather than system-wide.
Solution Approach 2:
The system uses automated parameter changes in popularity scores based on test results, reducing the need for manual evaluation of each completion. This automated parameter adjustment simplifies the overall system complexity by replacing manual analysis with algorithmic decision-making, while still maintaining high completion quality through data-driven insights.
Data Source
AI summary
Technology described herein identifies and demotes query completions which divert users from their original search intent. The identified query completions may represent informational requests that are amusing diversions for some users, but which are not relevant to the actual information of interest for many users. For example, the identified query completions may represent information requests that some users may find humorous or entertaining, but which are not relevant to actual information of interest for many other users. The identification of these query completions may allow alternative query completions to be provided that can more likely assist users in finding the information they actually seek.