Salesloft is the only AI-powered revenue platform that brings certainty to every single revenue customer interaction. The web-based platform handles all customer communications and guides go-to-market teams towards the right actions to yield the best outcomes: delighting customers, maximizing revenue, and increasing lifetime value.
When Salesloft started working with QA Wolf, each of their 25 squads had a manual tester to support the development process. Across the entire organization, the manual testers were responsible for more than 3,000 separate test cases. While the squads were diligent about testing their own focus areas, bugs would escape when a release impacted another area of the application.
As a continuous deployment, completing a comprehensive regression suite on every PR simply wasn’t practical. And with the growth trajectory of the company, doubling or tripling the number of manual testers wasn’t sustainable. So the leadership at Salesloft set internal deadlines to automate their regression testing and pursued several strategies to get there.
No-code tools were limiting and a burden to manage
Salesloft’s first attempt at automation was using no-code and low-code tools that purported to make testing easier. These tools, like any automation framework, create an enormous maintenance burden for the teams responsible for them. And ultimately, it became more work to maintain the test suite than to test things manually.
“The no-code tools just didn’t sit well with the squads. They were hard to maintain. We had lots of flaky tests. And we just didn’t have the capacity, the people, or the time to deal with that. What I love about QA Wolf is that you guys take care of everything from the test writing to the maintenance.”
—Ann Rumney, Staff QA Program Manager
Developer-driven testing left coverage gaps
When no/low-code tools didn’t deliver the results Salesloft was looking for, they turned to developer-driven test design and management. With their engineering talent, automation was relatively simple but test planning was a new skill and left gaps in each squad’s testing suite.
“They weren’t jazzed about it. I've always said you kinda have to have a QA brain and know how to actually think about testing; to think about the different variations and different edge cases that nobody else would think of. It makes a big difference to have QA people writing automation scripts.”
—Ann Rumney, Staff QA Program Manager
Manual testers struggled to test releases while learning automation
The team also tried to train their manual QAs in automation frameworks but the team wasn’t moving fast enough and the number of test cases was continuing to expand. The longer the team went without a sustainable automation strategy, the farther behind they got from their goals.
Salesloft estimates that it would take up to seven full-time QA engineers to provide the same service that QA Wolf provides today — to build, maintain, concurrently run, and respond to failures within strict SLAs. And it would have taken twice as many people to build the test suite in the time it took QA Wolf to do it.
"Hiring the number of SDETs we would need to get off the ground running would cost too much time and money."
—Ann Rumney, Staff QA Program Manager
During major upgrades like a new version of Rails, all 25 manual QA engineers would be pulled from their teams to go through 800 P0 workflows, which then delayed their squads’ ability to release causing additional delays across the organization. With QA Wolf, those tests can be completed in less than 45 minutes and they can be run on every PR.
“QA Wolf’s parallel run infrastructure lets us run the test suite so much faster than we could ever have run them on our own. I’ve been doing QA and test automation for 25 years and I’ve never seen anything like QA Wolf before — I think it’s going to serve a lot of people going forward.”
—Ann Rumney, Staff QA Program Manager