跳转至

Setting up site analytics

As with any other service offered on the web, understanding how your project documentation is actually used can be an essential success factor. Material for MkDocs natively integrates with Google Analytics and offers a customizable cookie consent and a feedback widget.

Configuration

Google Analytics

7.1.8 · Default: none

Material for MkDocs integrates with both, Google Analytics 4 and the now phasing out Universal Analytics. Depending on the given property prefix, add the following lines to mkdocs.yml:

extra:
  analytics:
    provider: google
    property: G-XXXXXXXXXX
extra:
  analytics:
    provider: google
    property: UA-XXXXXXXX-X
How to measure site search usage?

Besides page views and events, site search can be tracked to better understand how people use your documentation and what they expect to find. In order to enable site search tracking, the following steps are required:

  1. Go to your Google Analytics admin settings
  2. Select the property for the respective tracking code
  3. Select the data streams tab and click the corresponding URL
  4. Click the gear icon within the enhanced measurement section
  5. Ensure that site search is enabled
  1. Go to your Google Analytics admin settings
  2. Select the property for the respective tracking code
  3. Go to the view settings tab
  4. Scroll down and enable site search settings
  5. Set the query parameter to q

Was this page helpful?

8.4.0 · Default: none · Experimental

A simple feedback widget can be included at the bottom of each page, encouraging users to give instant feedback whether a page was helpful or not. Add the following lines to mkdocs.yml:

extra:
  analytics: # (1)!
    feedback:
      title: Was this page helpful?
      ratings:
        - icon: material/emoticon-happy-outline
          name: This page was helpful
          data: 1
          note: >-
            Thanks for your feedback!
        - icon: material/emoticon-sad-outline
          name: This page could be improved
          data: 0
          note: >- # (2)!
            Thanks for your feedback! Help us improve this page by
            using our <a href="..." target="_blank" rel="noopener">feedback form</a>.
  1. This feature is natively integrated with Google Analytics, which is why provider and property are also required. However, it's also possible to provide a custom feedback integration.

  2. You can add arbitrary HTML tags to the note which is shown after the user submitted the feedback, e.g. to link to a feedback form.

Both properties, title and ratings, are required. Note that it's allowed to define more than two ratings, e.g. to implement a 1-5 star rating. Since the feedback widget sends data to a third-party service, it is, of course, natively integrated with the cookie consent feature1.

How to visualize the collected feedback ratings?

To visualize feedback ratings you'll need to create a custom report with Google Analytics that will quickly show you the worst- and best-rated pages of your project documentation.

  1. Go to your Google Analytics dashboard

  2. Go to the configure page on the left hand menu, then select custom definitions

  3. Click the custom metrics tab and then create custom metrics, enter the following values:

    • Metric name: Page helpful
    • Description: Was this page helpful?
    • Event parameter: data
    • Unit of measurement: Standard
  4. Go to the explore page on the left hand menu, create a new blank exploration

  5. Configure the report as follows:

    • Dimensions: Add Event name and Page location
    • Metrics: Add Event count and Page helpful (the custom metric created in step 3)
    • Rows: Page location
    • Values: Drag in both Event count and Page helpful
    • Filters: Add a new filter for Event name / exactly matches / feedback

Delay in data availability

The report may take 24 hours or longer to begin displaying data

  1. Go to your Google Analytics dashboard
  2. Open the customization panel on the left and go to custom reports
  3. Create a new custom report and set a custom title and name
  4. Add Avg. Value and Total Events to metric group
  5. Add Event Label to dimension drilldown
  6. Add Event Category to filters and filter for the value feedback

Now, after you've saved the report and collected some feedback ratings, you'll have a list of all pages with the total number of ratings, and an average rating per page. This should help you identify pages that need to be improved:

feedback report

The following properties are available for each rating:

icon

Default: none · Required – This property must point to a valid icon path referencing any icon bundled with the theme, or the build will not succeed. Some popular combinations:

  • + material/emoticon-happy-outline + material/emoticon-sad-outline
  • + material/thumb-up-outline + material/thumb-down-outline
  • + material/heart + material/heart-broken
name

Default: none · Required – The value of this property is shown on user interaction (i.e. keyboard focus or mouse hover), explaining the meaning of the rating behind the icon.

data

Default: none · Required – The value of this property is sent as a data value with the custom event that is transmitted to Google Analytics2 (or any custom integration).

note

Default: none · Required – The value of this property is shown after the user selected the rating. It may contain arbitrary HTML tags, which is especially useful to ask the user to provide more detailed feedback for the current page through a form. It's also possible to pre-fill forms with the URL and title of the current page by using the following placeholders:

  • {url} – Page URL
  • {title} – Page title
https://github.com/.../issues/new/?title=[Feedback]+{title}+-+{url}

In this example, when clicking the link, the user is redirected to the "new issue" form of your repository, with a pre-filled title including the path of the current document, e.g.:

[Feedback] Setting up site analytics – /setup/setting-up-site-analytics/

An alternative to GitHub issues is Google Forms.

Usage

Hiding the feedback widget

The feedback widget can be hidden for a document with the front matter hide property. Add the following lines at the top of a Markdown file:

---
hide:
  - feedback
---

# Document title
...

Customization

Custom site analytics

In order to integrate another analytics service provider offering a JavaScript-based tracking solution, just follow the guide on theme extension and create a new partial in the overrides folder. The name of the partial is used to configure the custom integration via mkdocs.yml:

<script>
  /* Add custom analytics integration here, e.g. */
  var property = "{{ config.extra.analytics.property }}" // (1)!

  /* Wait for page to load and application to mount */
  document.addEventListener("DOMContentLoaded", function() {
    location$.subscribe(function(url) {
      /* Add custom page event tracking here */ // (2)!
    })
  })
</script>
  1. As an example, this variable receives the value set in mkdocs.yml, which is "foobar" for property.
  2. If you're using instant loading, you can use the location$ observable to listen for navigation events, which always emits the current URL.
extra:
  analytics:
    provider: custom
    property: foobar # (1)!
  1. You can add arbitrary key-value combinations to configure your custom integration. This is especially useful if you're sharing the custom integration across multiple repositories.

Custom site feedback

A custom feedback widget integration just needs to process the events that are generated by users interacting with the feedback widget with the help of some additional JavaScript:

var feedback = document.forms.feedback
feedback.addEventListener("submit", function(ev) {
  ev.preventDefault()

  /* Retrieve page and feedback value */
  var page = document.location.pathname
  var data = ev.submitter.getAttribute("data-md-value")

  /* Send feedback value */
  console.log(page, data)
})
extra_javascript:
  - javascripts/feedback.js

 


  1. If the user doesn't accept the analytics cookie, the feedback widget is not shown. 

  2. Note that for Google Analytics, the data value must be an integer.