DIY metrics for a wiki space dashboard

[#1]

Unique visitors. Page views. Time spent on page. Pages per visit. Entry and exit pages. Click path. All those cool web metrics that help people decide where their site needs improvement?

I don't have those. Yet.

But by being creative—and very, very stubborn, I mean determined—I have different metrics. Labels in use. Pages without labels. Stale pages. Heavily-edited pages. Pages created over time. Pages with most comments. Pages created by section by quarter. These numbers help me know what to pay attention to. This article is about how I went about getting them.

Because reasons, analytics add-ons and plug-ins haven't been available within our wiki instance to space admins like me and the average end user. This was disappointing at first, so the next obvious question was, "Well, what metrics CAN I get, with the add-ons I've got?"

To answer that question, I thoroughly investigated our current add-ons to find that sweet spot where meaningful numbers and available numbers overlapped. The results are now populating a "wiki dashboard" for one of the spaces I curate. I'm still improving the layout and some charts, but below are a few snapshots. The page is divided into three main sections: content, users and engagement.

Content

• Total pages added over time
• Section pages added over time
• New pages created by quarter
• Most edited pages (most versions)
• Pages last edited a long time ago
• Section pages average number of versions
• Number of labels in use
• Most popular labels
• Pages without labels

Users

• Top contributors
• Top labelers
• Unique page creators - Ideally, this number should increase over time as more individuals contribute to the space

Engagement

• Most commented pages
• Most liked pages (TBD)
• Top commenters
• Top page watchers

About each section / report

Each dashboard section contains three elements:

• The significance of the numbers (to me) (why bother to build the report?)
• The source of the numbers (mostly out-of-the-box Confluence macros and ServiceRocket's Reporting add-on). A couple sections combine a Reporting table and Confluence's Chart macro; I'll write a how-to on that soon. ServiceRocket's excellent support staff helped me out when I got stuck building a specific report.
• The chart, list or report itself.

Actionable data

What do I do with this information?

• I recognize and encourage top contributors
• I encourage people to contribute when new page growth is flat
• I review pages with comments to ensure questions don't go unanswered
• I add labels to unlabeled pages to ensure Content by Label macros can find them
• I monitor pages with many edits for quality control and make sure someone isn't struggling to edit a page
• Watch for duplicate or misspelled labels
• I review stale pages and get them updated or archived.

One of my favorite sections is "top page watchers." The number of people watching the ENTIRE SPACE increases often! In my mind, these people care so much about the wiki that they are willing to receive many, many emails detailing many, many edits.

"Most likes" is an elusive metric I'd really like to have. Neither the Content Report Table nor the Reporting suppliers can do this for me. For page views, we are testing the Tracking add-on but I'll have to wait for implementation in production. Until then, I will continue to explore how many insights I can gather with the available tools.

[Richard Bouchacourt]

Hello @Michelle Rau HP 

Useful article, I love questions. We are going further with questions !

I want to share with you my experience with KPI website analytics.

Several KPI like bounce rate could be view by two ways.
1/ Internet users came on a page and they found the informations.
2/ Internet users came and they don't found the informations.

How to analysis the data ? I crossed the bounce rate with the time spend on the page. If the spent time is very short, the page doesn't answer the user search query. Moreover, the time spent on the page is relative to the content (short/long, tutorial/article, people news/research paper, news website/e-commerce website).

Data are only data. I need to interpret data with humans behaviors, culture, literacy and informatic skill of the target Internet users.

Richard

[Michelle Rau HP]

@Richard Bouchacourt you are absolutely right. Data is just data until we apply our understanding to get real insights we can act upon!

[Azwandi [ServiceRocket]]

This is like a must-read article for those who value performance of their documentation.

Great piece, Michelle. Thank you for sharing. Love the details and your creativity in answering those questions.

[Michelle Rau HP]

Thank you so much @Azwandi [ServiceRocket] ! What nice words!

[neil briscombe]

"This article is about how I went about getting them." I can't find that in the article is there an update or part 2? Added a like for the principle though :)