Following up on Josh Clayton's post Filter out internal/agent views when reporting on Help Center article views.
You have one internal Category, a couple of internal Sections, and a total of ten internal articles (ie. articles that end users can't access; only agents).
It's fairly simple to obtain these thirteen IDs and create a URL-based exclusion filter in Google Analytics or Data studio. But what if we scale this example to the hundreds of articles, sections, and cateogies?
No permalink structure
If we could create a custom URL structure like, for example, domain.com/hc/en-us/categoryID/sectionID/articleID, this problem could be solved.
Suppose all my internal content is under category ID 123465789... if this ID could be preserved in section and article URLs, then '123456789' it would be the only expression I'd need to exclude in Analytics.
However, Guide ignores content hierarchy on its URLs. We only get one ID: either the category's, the section's, or the article's. Consequently, the only solution is to make sure we have a full list of all the internal IDs and exclude them in Analytics.
Side inconvenience: if you use the Google Analytics add-on for Google Sheets, it's going to be a mess, since there is a limit to how long your filter expression is... too many internal URLs? You'll have to split these into multiple individual reports:
Detect user role and pass that information to Analytics
So there are several layers of "user vs content visibility" that we could include/exclude:
- Any user visiting any content
- Any user, visiting public content
- End user, visiting public content
- Agent, visiting any content
- Agent, visiting public content
- Agent, visiting internal content
When we install Google Analytics, we're measuring #1. If we add URL-based filters as exclusions in Analytics, we get all user activity related to public articles, i.e. we can get #2 and #6.
However, I'd still need to exclude agent activity on public article IDs (#5).
For example, the knowledge base team who are constantly double-checking keywords vs search results. The impact are probable incorrect/incomplete assumptions, and conclusions, in our data analysis.
So maybe there is a simple way to do a role identification and restriction, and I lack the knowledge.
Please sign in to leave a comment.