Question
How is the Related articles list in the Help Center populated? Can I influence what articles are returned?
Answer
You cannot influence what articles are displayed in the Related articles list.
Articles in this list are returned based on a relevancy score between articles. This score is calculated via the following method:
- For each user, we keep a list of their recently viewed articles.
- When a user visits an article, we record the most recently viewed article as being related to the currently viewed article.
- Over time, each article will be related to a number of other articles. The more times an article is marked as related to another article, the higher the related score is. The score and therefore the results will get better over time.
For example:
- User visits Article A.
- From Article A, the user visits Article B.
- Article A's relevant score in relation to Article B is increased.
- Another user visits Article B and Article A is suggested as a related article.
For more information, see our developer documentation: Recent articles helper.
18 Comments
So... basically, "Recently Viewed" ultimately becomes "Related"?
And the content creators cannot provide any guidance as to what is actually related?
Do we have any feature or Ability to set ‘related FAQs’ against each FAQ. manually?
Any suggestions please welcome..
Thanks for your question!
I asked around with some folks here and their response is that the tool doesn't offer that kind of functionality; the actions are all organic based on the formula outlined in the article.
However, it is possible for you to write something custom to allow you to do what you want to do.
This functionality would be extremely appreciated if it was added. Right now I'm adding a "See also:" link at the bottom of my FAQ questions, but this really should be a smarter feature that allows for simply checking a box next to articles in a list, possibly with another checkbox for whether we want that link to open in a new window.
As it stands now, the Recently Viewed and Related Articles are just redundant variations on the same theme and aren't particularly helpful. Site visitors won't care, but they probably won't use those links all that much, either.
I can see that you use the term "user", but does it mean that the feature only calculates related articles based on logged in users or will it be all visitors?
I think this feedback is what we want?
@Karsten - yes, it is just logged in users.
misleading, not a helpful feature.
when i read the title related article, i am expecting to show all articles have the same label.
We noticed a lot of crawl errors in our google search console from the related links. Can these be generated to use actual urls to other help centre pages so that we do not received errors for the short term links?
Hey Joe,
If you're receiving crawl errors we can certainly dig into that more with you and see if we can figure out where the issue is coming from. Unfortunately there's no way to change the generated URL's.
I'm going to create a ticket on your behalf and pass this over to our Customer Advocacy team so they can assist further. You'll receive a follow-up email shortly stating your ticket has been created. Can you reply back to that email with any examples you have available so we can take a look?
Thanks!
We've been working with Envoy (partners of Zendesk) and they thought that titles of articles and tagging may influence "related articles". Is this true? Can we use that to help the algorithm of "related articles"?
Hi guys,
I was planning to add an extra section (similar to Related articles) that populates itself using the stats of the kind of tickets open by the current user logged in our KB.
Example: if user opens 5 tickets in the last 30 days regarding VPN issues, 3 regarding Bluejeans, 4 regarding Outlook, etc.... when the user logs into our KB, this "smart section" will show articles related to the user tickets history.
is there a way to make the call?
Thanks
Hey Cristobal -
That's an interesting idea. It would definitely require some custom code; I've shared your post with some customization experts, so hopefully one of them will jump in with some suggestions soon.
Hi Nicole,
I am happy to run custom code as we have added quite a lot already. I just need to find out the call to use to request the info from Support API I guess?
Thank you
Ok. How are you tracking what users file tickets about?
That's the key question, if users login to our Knowledge base, they can see their own history of tickets, so that can be the call.
Another option, from Support perspective, if I search for a user, I can see all the tickets open by this specific user...so that can be the alternative call.
Every ticket open by the user has a category, so my idea is to retrieve those tickets by category, and the top 5 categories used in the last month for example, find the article related and show these articles in this "smart section" based in the history of tickets open.
Probably is not an easy call and needs to be done using multiple calls, but definitely this will be helpful for the users.
My idea obviously is to make it more user related, so each user can see their own list of articles in this section.
If that is not possible, then maybe I have to change the perspective and do it more generic, so taking the overall tickets open in the last month (as an example), organise it by categories and the top 5 categories used, find the articles related and show the articles in this smart section.
But I prefer the first option so all the users have related content under their own experience.
So the key here is to know how to make the API requests and include them in the Javascript file.
Thank you
Got it. So it sounds like you have fields/tags that are tracking the category the tickets fall into and then corresponding fields that your articles are classified by.
We'll include your post in our digest this week to see if any other users have suggestions on how to do this.
Otherwise, you could reach out to our Professional Services team to see if this is something they could help you build.
Please sign in to leave a comment.