Plan requirement
| Subscription | Suite Professional or higher, Knowledge Professional or higher |
| Access | Admin |
Views and votes are the numbers people watch. Two others tell you far more about whether your knowledge base works.
The two that matter most
Searches that returned nothing
Every one is somebody who came looking and left without an answer. It is a list of articles you should write, in your customers' own words.
No other report gives you that so directly, and almost nobody looks at it.
Tickets about subjects you have articles for
Harder to measure and worth the effort. Somebody raising a ticket about something you documented means the article was not found, or was not good enough. Either is fixable.
The ones people over-read
- Views. They tell you what is popular, which you probably knew. A high-traffic article is worth keeping current; that is most of what views are for.
- Votes. Useful for spotting outliers, misleading as a score. Articles about annoying problems get poor votes regardless of quality.
The combination worth building
Most-read articles with the worst ratings. That is the shortest list of the highest-value fixes, and it is one report.
Explore does offer a net figure, and that is the one to avoid. Net article votes is upvotes minus downvotes, and a net score hides the article that eight hundred people read and forty disliked behind the one that three people read and one disliked. Build the report from views against Article downvotes instead. Watch for Net article votes - all time as well: it ignores your time filter, so a dashboard that looks filtered is not.
A third worth adding
Searches with no clicks: somebody searched, got results, and opened none of them. That is a different failure from an empty result list. The words were found and the titles did not look like the answer, which is a rewriting job rather than a writing one.
The number that is not in Explore
Whether ticket volume falls after you publish something. That is the actual purpose, and connecting the two takes deliberate effort. It is also the only argument that convinces anyone to fund knowledge work.
One practical limit before you promise anybody a trend: self-service data starts on 5 March 2025. Anything you want to compare against has to sit after that date.
Comments
0 comments
Article is closed for comments.