Page MenuHomePhabricator

Generate meta description tags to article pages for better search engine indexing
Open, Needs TriagePublic

Description

We don't currently generate meta description tags for our article, which means we're trusting fallible search engines to find a snippet in the article. While this task should be trivial -- just grab the first 150 characters of the article! -- it doesn't appear to be e.g. the current enwiki entry for Barack Obama:

Screen Shot 2015-07-20 at 2.58.10 PM.png (212×1 px, 70 KB)

And it gets particularly hard given our tendency toward IPA codes e.g.:

Screen Shot 2015-07-20 at 3.08.24 PM.png (178×1 px, 61 KB)

Google cleans these up in its extracts of the Wikipedia article in the Knowledge Graph box which makes our odd snippet look even odder:

Screen Shot 2015-07-20 at 3.12.27 PM.png (224×872 px, 41 KB)

Screen Shot 2015-07-20 at 3.10.37 PM.png (290×862 px, 56 KB)

Should we generate meta description tags and add to the head of pages? Been discussed before (e.g. https://phabricator.wikimedia.org/T14196 from 2007) but feels like an easy opportunity to improve the way many users first see a Wikipedia article (in search results).

Event Timeline

Stu raised the priority of this task from to Needs Triage.
Stu updated the task description. (Show Details)
Stu added a project: SEO.
Stu subscribed.

According to Lighthouse Best Practices, document should have a meta description. See here. The meta description should include the page title or the content of the first header.

Is there a plan to implement the meta description in the near future?

@Fuzzy: There is not (in general this applies to any unassigned task). Code contributions are welcome.