What criteria should not be used when filtering results in a saved question source?

The filtering of results in a saved question source is essential for tailoring the data to meet specific needs. When considering the options, filtering by related values, filtering by date, and filtering by computer groups are all valid criteria that can provide meaningful insights based on various dimensions of the data.

Filtering by related values allows users to focus on specific attributes that are related or of interest, helping to narrow down the data set. Filtering by date enables the analysis of data within specific time frames, which is crucial for tracking changes and trends over time. Filtering by computer groups is also a useful approach as it allows for segmentation of data based on device classifications, providing deeper insights into specific segments of the network.

Filtering by percentage does not align with established criteria typically used in saved question sources because percentages often require a context or a base for what is being measured, which might not be available in the data source. Furthermore, percentages may not be directly applicable or meaningful in filtering results, as they usually serve as a way to represent data rather than a method for segmenting or querying it.

Therefore, while the other criteria serve practical purposes in refining results, filtering specifically by percentage does not provide a viable means of filtering in a saved question source.