May 2020

Goodbye, Loadster Workbench

Loadster Workbench started as a personal side project, and I first shipped it to the general public in early 2012.

Despite my bumbling incompetence around sales and marketing at the time (only slightly improved in recent years), it managed to eek out its first sale in February 2012, just a couple weeks after release! I don’t recall how that first brave soul found the site, but I distinctly remember where I was (out to dinner at a local restaurant) when the SMS arrived saying someone had paid $399 with their credit card.

In those days, Loadster Workbench generated all the load itself with no help from external engines, and was capable of maybe a couple thousand v-users… depending on all the typical hardware and software constraints of the system it was running on. It supported Windows only at first but soon shipped for Mac also. The desktop app was written in Java, using the SWT library to wrap native widgets, and bundled as an EXE on Windows and an app bundle on the Mac. Because of all this, it was able to pass for a native app better than most Java desktop applications tend to. It even had a fancy installer on Windows and the drag-and-drop DMG package on Mac.

Like most people I suffer from a fair amount of imposter syndrome, and I only gradually got used to the Founder title. I still don’t call myself a CEO. Yet sales of Loadster Workbench continued to grow in fits and starts and it got to the point where it was making me a decent chunk of side income, while actually being useful for hundreds of actual people at actual companies. I met people for whom a big part of their job was using Loadster to run load tests and improve performance in their web apps! Even today, it sort of boggles the mind.

As more and more people started using it, I noticed a few things about the business of shipping and supporting a desktop application. For one thing, getting to the bottom of weird platform-specific bugs is a pain. Customers had quite a few different OS versions to begin with, making problems difficult to diagnose, and some of them also had complicated comorbidities such as antivirus software and outbound proxy servers. Some users didn’t even have admin rights on their own box! And while the widget toolkits I was using made a Java desktop app feasible, they were a bit long in the tooth and lacked the energy of the web stack, which is relentlessly getting more powerful and feature rich all the time.

There have also been obvious, so-big-you-can’t-miss-them shifts in the industry. Back in 2012, most teams still had in-house test environments and many even had self-hosted production environments, so a cloud-only tool would have ignored a large swathe of the market. But in the subsequent years, of course, nearly everyone has moved to the cloud. So there’s less and less natural demand for a desktop application.

In light of these trends and observations, I started repackaging Loadster as a web-based SaaS in early 2018, and it quickly started getting adoption at a faster clip than the Workbench ever did. Onboarding was simpler since you never have to leave your browser. Conversion rates were better. Team collaboration was easier. And building new UI in a Vue.js SPA is a breeze compared to the herculean task of building a rich and engaging interface with Java/SWT. During that time period, we added a couple more team members and took a modest amount of funding, but still remain a small and independent software team.

Fast forward to early 2020, and the vast majority of our customers were happily using the web version. A few loyal stalwarts were still using Loadster Workbench, but it became apparent that we can do best by all our customers if we go all-in on the web platform and help the remaining Loadster Workbench users migrate over. We set an “End of Life” date for Loadster Workbench as May 1st, 2020, which is now in the rear view mirror.

I’m excited about the accelerated progress now that we’re 100% dedicated to the web platform, and even more excited about some other things currently in the works, that just wouldn’t be possible with a desktop application. We’ll announce them here soon!

– Andy Hawkes, Founder at Loadster

View the Changelog