This proposal is a larger suggestion that is out of scope for the Community Tech team. Participants are welcome to vote on it, but please note that regardless of popularity, there is no guarantee this proposal will be implemented. Supporting the idea helps communicate its urgency to the broader movement.
Improve pageviews to show which parts of the article the user is reading
Problem: The current pageview analytics are not specific enough to direct an editor's attention to the portion of an article that is most in need of being improved.
Proposed solution: The proposed solution is analogous to the YouTube engagement analytic that shows, statistically, at what time in a video viewers are leaving. However, the proposed solution for a Wikipedia article needs to account for multiple entry points into the article, as opposed to YouTube viewing which usually starts at the beginning of the video. Therefore, the proposed solution is to provide article editors with data showing how much time readers spend on each rendered section of the article, starting at the reader's entry point into the article.
Who would benefit: The article editors would benefit from the feedback and the article readers would benefit from articles that have had their worst parts improved.
Good idea. Such analytics might also give an idea of which parts of an article can be deleted! Similar to the grammar-check proposal. In general Wikipedia needs to get with the times and make more use of automated tools. They represent free productivity, there for the taking. That should be a priority for a project built on volunteer labor. --Rollo (talk) 19:17, 16 January 2022 (UTC)[reply]
I agree that it could help with deletions. I think deletions are the most difficult edits because of the risk of discouraging volunteers from making future contributions. However, deletions often improve the article and having some data to help rationalize deletions would make deletions a little less difficult. --Gj7 (talk) 15:34, 17 January 2022 (UTC)[reply]
This is very much out of scope for Community Tech, but we're going to put in with our Larger suggestions category so it can still get some attention. Without having ran this by the Analytics team (who maintains the pageviews pipeline), I suspect there are many concerns with this. One is that we'd basically be monitoring what parts of the page the reader is viewing, which is something perhaps not everyone is comfortable with (even if anonymized). It could also be of performance concern for mobile readers, given we'd have to continually make roundtrips to the server even though the user hasn't done anything beyond scrolling. It also would be subject to many false positives and could send editors the wrong signal. For instance, a reader may view the table of contents for an article on a music artist and jump straight to the Discography section. This doesn't mean there's anything wrong with the content in between; they just want the discography and nothing else. In my opinion an encyclopedic article isn't that comparable to a YouTube video in this regard. MusikAnimal (WMF) (talk) 23:45, 18 January 2022 (UTC)[reply]
The rewording of the wish changed its nature. The original wish was not to track a reader's movement through an article as the current wording suggests. The problem, as I see it, is that article editors are blind to the statistical patterns of reader access: including time and order. In the case of your example, if 90% of the readers of that article on a music artist jump straight to the discography section, would that not be useful information for article editors to help reorganize the article, for example by putting the discography section first?
Oppose: this information would be both informative and motivating to many editors in many situations, but this is outweighed by privacy (and possibly performance) concerns. We should be proud about not collecting this sort of intrusive analytic data. — Bilorv (talk) 20:35, 28 January 2022 (UTC)[reply]
Oppose: There would be a performance hit from such analytics. Wikipedia doesn't need this anyway - let's strive to have excellent articles, not just excellent sections. --Šedý (talk) 11:24, 29 January 2022 (UTC)[reply]
Support Obviously it would need to be implemented in a privacy-conscious way, but I think more information could help with article decisions Ph03n1x77 (talk) 07:07, 5 February 2022 (UTC)[reply]