The screenshot tool defaults to a screen layout instead of a full page test. I find it a bit cumbersome that I can't have it run a full screenshot as my default. Every time, I have to select the full screen, then restart its captures, which seems a waste of time and energy.  This is, admittedly, a minor complaint.

Web user interface is "heavy" and slow to load. Some browser configurations do not work at all times. Account limits do not allow you to queue jobs - you must wait to submit until jobs complete. Could use improvements in ability to tag on create screen.

Sometimes the testing is slow. The loading time could be improved for faster performance.

This solution would benefit from faster testing and support for more devices.

Being able to test on real devices via the virtual connection is wonderful, but it can cause some lag and load time issues while testing. This is one of the few concerns our team has had while trying to accomplish our testing. Going forward, I'm sure improvements can be made to help alleviate some of those concerns.

Expansion on automation and quicker testing is something that would greatly benefit our business needs. Smartbear has already implemented some automation on the platform that can be used and improved to fulfill our needs as well as many other users. 

A few intermediary pricing options for small QA teams would be nice, e.g., unlimited screenshots, "as you need it" parallel tests, etc.

We had some issues with the onboarding process and the cloud conductivity could improve.

I have had quite a few issues trying to use a virtual machine to test our application on. Sometimes it will freeze or not respond. I cannot log in to our application because the VM does not always respond to keyboard functions.

I would also like to see more real devices in this list of those to choose from. I have had some of our customers come to us with bugs that are specific to a particular mobile device and I could not recreate the issue since we do not have the particular device and it is not on crossbrowser Galaxy S9 .Id like to see Galaxy Tab S4/ E lite added.  

A wider range of physical devices with more browser versions in the Selenium Grid would be great to insure users with out-of-date devices are able to interact with our sites.

I have experienced some lagging issues, and it does not seem like all of the testing environments are configured the same. For example, I have had to figure out how to get to Safari or whatever program I needed. If these issues were resolved, it would be a huge improvement. 

• The speed connection in mobile devices could be improved, because sometimes the load time is uncertain. 
• The future possibility to do a web inspector to debug any site, on any device in any OS, would be a really nice feature for the service. This means not just with the web inspector included from each web navigator (Firefox, Chrome, IE, etc.).

It would be useful if we can run the live-testing test cases on multiple platforms at the same time, instead of waiting for one session to finish.

It would be helpful to have more help notes or videos to show how to build an automation framework.

I had some problems with the local connection, but there is someone that can help us always on customer support chat. They solved my problem.

The "Getting Started" documentation for Selenium testing could be improved. I was able to figure it out fairly easily, but I imagine some people might have trouble.

Elements of 'real' mobile/tablet testing could be sped up, as I still sometimes experience lag, which no one loves.

The five minute timeouts can cause irritation if you have just popped away to consult some supporting documentation.

• The speed to connect to mobile devices needs improvement, and sometimes the connection fails. 
• Possibility to use web inspector in all devices (it is not available for OS devices). 
• Also, easy access to run local files. I need to run a CSS file before loading the website, and sometimes, it does not work.
  

A problem that we are facing quite often is related to the network connection. Tests can fail if the remote CrossBrowserTesting's VM has connection problems. This happens mostly with browsers of Internet Explorer family which work on Windows OS. I would like CrossBrowserTesting's engineers to deal with this issue.

There should be more detailed training on CrossBrowserTesting.                                

Automated testing could be improved.

Test management feature where tests could be scripted.

Data-driven testing could be added.