Web analytics

These are usually the easiest numbers to get: plug in a tool like Plausible, and start seeing page hits, bounce rate, entry and exit pages, and so on.

The problem with analytics in docs

There is a problem with using analytics in docs sites: the numbers don't mean what they normally mean. Let's look at an example.

Page A (which documents Feature A) gets loads of hits. On a blog or shop site, that is almost always good news. On a docs site, it could mean many things:

🟢 This is a heavily used feature, so the docs are also heavily used. All is well.
🟡 This is an important feature to potential customers, so lots of them visit when evaluating. All is well, but it'd be nice to know this specific audience info, and maybe add extra material for decision makers here.
🟠 Feature A is confusing or counterintuitive. The docs are fine, but you might want to look at the feature itself.
🔴 Page A has a very similar title and description to Page B, and people are landing here by mistake. All is not well. We might try to determine if this is the case using bounce rate, but bounce rate can also mean people quickly found what they want and exited. To determine this effectively, you'd need a detailed user path: who gets here from search, and who goes straight from this page to searching again.

This is just one example. You can do a similar exercise for other common metrics, such as bounce rate.

The other problem with analytics (everywhere)

People are increasingly privacy-conscious. They might tolerate tracking on the docs site itself, but would they tolerate tracking within the app, which is probably needed to track the full user path and obtain certain success metrics?

This means you need to be thoughtful about your data gathering. It may also mean you can't (or even shouldn't) get all the data you'd like.

Recommendation

When setting up your analytics, get clear first on what you're measuring and what it means.

Some techniques:

• Combine multiple metrics.
• Differentiate by content type: for example, you might use time on page as a metric to measure tutorial content, but find it less useful for reference docs.
• Focus on signals of user success: if you can connect your docs analytics to your product analytics, you can track users as they move between the two, and see if docs improve user success with a particular feature.
• Use privacy-respecting analytics tools such as Plausible, instead of Google Analytics. This may make privacy-conscious users feel better about the data gathering, and should also lead to more accurate data.