Wiki Life: Why we don’t use Ratings on TechNet Wiki

Welcome everyone to todays Wiki Life.

Todays topic will be ratings. Ratings can be a nice tool to measure the usefulness/quality of content. This blog uses them for example, people can rate the blog entries, giving feedback to the authors what topics they liked and in that way influence what topics or style of presentation for content is used in the future. Forums have something similar, there you can vote a posting helpful, indicating again to the author if the content did help.

Sounds good so far, so why aren’t ratings used for the wiki? One of the differences between blogs/forums and the wiki is the dynamic of content. While writing a blog or forum post, you usually finalize the content. Of course you can edit it later on as an author, but you usually won’t modify it a lot, maybe spelling errors, broken links, or a note you believe should be added. In blogs, if you find the whole entry should be updated, you’d propably rather do a follow up blog post, or write a comment to your blog. Using the forum, you’d most likely simply reply to the ongoing discussion/conversation. The original posts are rather static content.

There is the big difference to the wiki. The entries on a wiki are ideally not static, but very much dynamic. Of course there are also metrics on the wiki that hint at the quality of an article. Not all metrics are visible out of the box, like the number of views an article has (the original author gets an indication with milestone archievements), how many other articles link to it (not visible right now to my knowledge, but who knows, might be something Peter collects over time, see his work here). One metric is visible however, and that’s the number of revisions, which also gives us the reason why “simple” ratings don’t work well with wiki articles. Instead of rating the article, you would basically rate the revision you see at that moment. Assume you would rate an article as being not so good content. Now someone comes along and improves the article (or you do revise the article yourself): the old revision described the problem, and the addition in the new revision also shows how to work around the problem. The article now is better than the rating it recieved in the past. This can happen the other way around as well, say there is a article describing a fix to a problem with a product. However, the product gets a service pack, and now the described solution would cause more harm than it helps. The good thing about TechNet Wiki is that eventually the articles get modified by the community. The ratings however wouldn’t change. So that’s why there are no ratings used for wiki articles.

There is of course still the option to give feedback to an article. You can comment on them, or you can edit/add to them. And maybe some day in the future, more metrics will be visible, too. If you have ideas on this topic, you can always suggest them at or start a discussion on the wiki forum over here.


Comments (8)

  1. I appreciate the thought you put into explaining this situation. Thank you!

  2. Thanks for sharing this insight.

    I know it has already been requested many times in a number of places, however it would be nice to see the wiki page views displayed on each page. This would serve as an indication to the author, editors and the community at large the popularity and by extension the utility of the wiki post. This can also serve as an indication to others the need for other wiki posts like it and/or a wiki post with content related to it. My two cents.


  3. We used to have rartings on TechNet Wiki and removed them for this reason.

    An even more extreme example is that someone would stub out an article, then someone would land on it, expecting a full article. Because it is just a stub, the reader is disappointed and gives it one star out of 5. Then someone else comes in and writes the article, making it awesome. An awesome article now has one star that was intended for the Stub version only.

    Thanks Florian!

  4. Thanks for explaining why we don't have ratings on wiki 🙂

  5. I agree with Ed, there are many articles with only one star. Actualy MSFT should automise this .. As exemple

    100 views = 1 star

    250 views = 3 starts

    1500 views and 20 edits = 4 stars


    So we can be sure that the ranking is "reel"


  6. Gokan, Good thoughts. Thanks!

  7. Great explanation and agree that TechNet Articles should not be rated like blogs or questions on the forums, however I think it would be interesting to have present the number of views the article have.