Sometimes you need to put something down only to pick it up later. Case in point my golf swing. When I shank a ball or duck hook it, the best thing to do is step back from the range, collect my thoughts and think about what I’m doing wrong. The same logic applies to the job. A year ago I left Velocity with the intention of using WebPageTest in our lab. I wrote up a nice little blog, but I failed to suggest why we should use the tool. I had no business case, whatsoever.
Turn the clock forward 1 year. We haven’t touched WebPageTest in probably 9 months or so. Now I’m a little more prepared to give a few business cases and reasons that we could use WebPageTest. I have (4) initial ideas listed below. I’m curious if other members of the team can think of other “business cases” for us to use WebPageTest, that would be awesome.
- Continuous Integration on critical use cases
- Data Variance on a Use Case
- Impact of Customization on the UI
- Synthetic Testing in Managed Hosting
- Geo-location testing for Managed Hosting
I could see some awesome value props in the areas above with the use of WebPageTest. I know we want browser testing for performance. We have been teetering with Selenium for ever, but haven’t embraced it like we would like to. WebPageTest could quite possibly be a replacement to Selenium. It would have a ton of value because of a couple of reasons:
1) We could share videos with the UX team and they would have meaningful data/visualizations to look at.
2) The scripting is a ton easier then Selenium and time to market would be a lot faster.
3) We could totally harvest the HAR data and do some wicked analytics inside of Galileo
There are couple of cool things about WebPageTest that we could take advantage of. First, we run our own private instance. Second, we can integrate via Fusion with all of their APIs. This would simplify the execution of our automated WebPageTest scripts (once we build them). Third, we can take the HAR file from the WebPageTest results and import/integrate with dynaTrace (BooYah!). This would give us the ability to analyze the data from our tests.
I think if we have some repeatable success with this, maybe we could convince MH to start considering using their own private instance of WebPageTest. That would be fantastic in my opinion…