Retiring low use articles systematically

Steve Allison
Tera Contributor

Hello, looking for any experiences with automatically retiring low use articles.

  • how did your org define low use? (views, edits, ratings, etc)
  • SN changes article status if a KB exceeds the annual review date- seems it could also retire them if low usage at the 12 month mark
  • opinions/best practices inputs welcomed

Thanks all

4 REPLIES 4

Sandeep Rajput
Tera Patron
Tera Patron

@Steve Allison It varies from organisation to organisation however, we designed a policy to retire articles on the basis of ratings and views. Articles were monitored for a specific duration and within this duration if the views and ratings for the articles were below certain thresholds we used to retire such articles automatically. Author were informed via notifications about the automated retiring of the article.

 

Hope this helps.  

John235
Tera Contributor

Hi, Steve...

We do not automatically retire KB articles.  We ensure the owner group is notified of a metric that is about to be exceeded.  Such as the annual review, no views within 365 days, etc.  

The owner group usually takes corrective measures to ensure the article is still relevant and accurate.  In the rare case that the owner group does not respond, the Process Owner may contact the owner group manager directly.

It is important to ensure the article is available with the correct information.  Not to automatically retire it without proper review.

Hope this helps,

John

 

Mike Van Vooren
Kilo Guru

@Steve Allison - We do use an automated retirement job since we do not use the yearly review.  We did not want this job to be overly disruptive or agressive, so the criteria are very loose.  We say to Reitre any article that has no Uses AND no Views over the last 6 months.  But we have some exceptions as well, like excluding End User-facing articles, exclude non-English articles were auto translated, and exclude articles in special Knowledge Domain groups (ie: Ownership Groups).  We may adjust this criteria over time, but for now this helps for now.

David Kay
Mega Guru

We have found that the best indicators for automatic archival are (a) Case / Incident attaches and (b) Self-Service page views.  The specific parameters you pick vary quite a bit based on the nature of the business--different product families in the same company often use different ones.  But an example would be, no Case / Incident attaches in 6 months, and fewer than 10 self-service PVs in the last year.

To figure out *your* parameters, make your best guess and run a report showing what you would be autoarchiving.  Then get with an SME / knowledge domain expert and ask them "Is this good? Can we archive more?  Have we cut too close to the bone?"  In other words, experiment with your parameters before archiving a single document.