Skip to content

Commit

Permalink
Readme edits
Browse files Browse the repository at this point in the history
  • Loading branch information
kaedroho committed Jul 20, 2021
1 parent 9f8ca35 commit 9959f85
Show file tree
Hide file tree
Showing 4 changed files with 31 additions and 13 deletions.
6 changes: 3 additions & 3 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -24,14 +24,14 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
- The tracking code has been moved into JavaScript to reduce the number of false-positives caused by bots
- The Cloudflare worker has been massively simplified

### Upgrading from, 0.2.2
### Upgrading from 0.2.2

Following the [installation guide](https://github.com/torchbox/wagtail-ab-testing/#installation), add the new URL pattern into your ``urls.py`` (note this is a separate URL to the one used by the old Cloudflare worker).
And add the tracking script HTML tag to your base template

#### If you use the CloudFlare worker
#### If you use the Cloudflare worker

Before deploying the update to 0.3, yoou must first update the code to match the [latest version in the readme](https://github.com/torchbox/wagtail-ab-testing/#running-ab-tests-on-a-site-that-uses-cloudflare-caching). Note that your A/B tests will temporarily stop working but this shouldn't have any other impact on your site. They should start working again once you've deployed 0.3.
Before deploying the update to 0.3, you must first update the code to match the [latest version in the readme](https://github.com/torchbox/wagtail-ab-testing/#running-ab-tests-on-a-site-that-uses-cloudflare-caching). Note that your A/B tests will temporarily stop working, but this shouldn't have any other impact on your site. They should start working again once you've deployed 0.3.

Remove the following URL pattern from your ``urls.py`` (note this is not the same pattern we added in the previous step!):

Expand Down
38 changes: 28 additions & 10 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -6,22 +6,40 @@ Key features:

- Create an A/B test on any page from within Wagtail
- Tests using page revisions (no need to create separate pages for the variants)
- Prevents users from editing the page while a test is in progress
- Calculates confidence using a Pearson's chi squared test

![Screenshot of Wagtail A/B Testing](/screenshot.png)
- It prevents users from editing the page while a test is in progress
- Calculates confidence using a Pearson's chi-squared test

## Usage

Any user with the "Create A/B test" permission (given to admins and moderators by default) should be able to create an A/B test by clicking "Save and create A/B test" from the page's action menu.
### Creating an A/B test

Any user with the "Create A/B test" permission can create an A/B test by clicking "Save and create A/B test" from the page's action menu.

The first page shows the user the difference between the content in the latest draft against the live version of the page.
This allows them to check what changes on the page are going to be tested.

Once they've confirmed that, the user is taken to a form to insert the test name/hypothesis, select a goal, and sample size.

![Screenshot of Wagtail A/B Testing create page](/screenshot-create.png)

### Monitoring test progress

While the test is running, the page's edit view gets replaced with a dashboard showing the current test progress.
Users cannot edit the page until the test is completed or cancelled.

Any user with permission to publish the page can start, pause, resume or end A/B tests on that page.

![Screenshot of Wagtail A/B Testing](/screenshot.png)

This will show a diff of the content in the latest draft against the current live version of the page (as these will be tested). Once they've confirmed that, the user is taken to a form to insert the test name/hypothesis, select a goal, and a sample size.
### Finishing the test

While the test is running, user's can pause/restart and end the test at any time. While the test is running/paused, nobody can edit the page as this may affect the accuracy of the test results. If you absolutely must edit the page while a test is running, you should cancel the test, edit the page, and start a new test.
The test stops automatically when the number of participants reaches the sample size.
Based on the results shown, a user must decide whether to publish the new changes or revert to the old version of the page.

When the number of participants reaches the desired sample size, the test is stopped and the results are displayed to the user. At this point, the user may choose to publish the latest draft (aka the "variant") or revert back to the existing live version (aka the "control").
Once they've chosen, the page edit view returns to normal.
The results from this A/B test remain accessible under the A/B testing tab or from the A/B testing report.

Wagtail A/B testing will also tell the user if one of the results is a clear winner or if there is no clear winner, this is worked out using a Chi Squared test that takes into account the total sample size (higher sample size means it will accept a smaller difference as a clear result) as well as the conversion rates.
![Screenshot of Wagtail A/B Testing](/screenshot-finish.png)

## Installation

Expand Down Expand Up @@ -70,7 +88,7 @@ Each A/B test has a goal that is measured after a user visits the page that the
The goal is defined by a destination page and and event type. For example, if the A/B test needs to measure how a change on the page affects the number of users who go on to submit a "Contact us" form, then the 'destination page' would be the "Contact us" page and the 'event type' would be "Submit form".

Out of the box, the only 'event type' that Wagtail A/B testing supports is visiting the destination page.
If you need to measure something else (such as submitting a form, purchasing someting, or just clicking a link), you can implement a custom 'event type'.
If you need to measure something else (such as submitting a form, purchasing something, or just clicking a link), you can implement a custom 'event type'.

### Implementing a custom goal event type

Expand Down
Binary file added screenshot-create.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added screenshot-finish.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.

0 comments on commit 9959f85

Please sign in to comment.