Improving WordPress.org updates through visual testing

While hosting companies added auto-updates for WordPress core and some for themes & plugins, when we look at charts on WordPress.org, very few users actually update their products, why? Cause in 70% of the cases their site is affected.

Why Visual Testing?

Websites are visual things

While code is also important for security reasons and much more, if you site breaks or looks like crap, you don't care too much.

Spot hard to detect changes

After analysing 100+ theme updates most of our changes can be easily overlooked, 10px extra spacing here or there, some left padding for an HTML item or some mobile only change, all very small but important enough.

No other way

For lots of possible problems there is no other way, while there are lots of code analysis tools is really hard to know if some piece of code will affect a user site or a page.

How EyePatch can help?

More reliable updates from authors

Theme & plugin authors can be informed when an update introduced visual changes and the update can require manual input if the difference is too big, this would automatically stop the update to go to users sites.


We test theme updates for example for compatibility with 3 major plugins as well, custom testing data can be configured.

More informed data to the users

Using the same technology we can detect if a user has pending updates and use that data to run a visual test, for example the user have a pending update for Hestia 2.0 from Hestia 1.4.5, we run a visual test and inform him.

User can now take an informed decision, he can check the report and wait for an update or hit update and pay attention to potential changes to fix them asap.

How it works?

No user tracking required

Just by looking at the update request that is already sent, and a standard and complete testing data, we create two separate instances, screenshot and compare them while returning the results.

Is not close to a staging website, however it can still dramatically improve the overall user experience.