Possible Finance offers short-term loans and credit to low-income earners who have been left out of the traditional financial system. Their proprietary underwriting model lets them offer flexible, non-predatory loan products where their incentives are aligned with their customers.
As the team at Possible released new features and services, bugs were starting to turn up in unpredictable areas of the application and eroding the trust of new and existing customers. For financial services companies like Possible, customer trust is as good as currency. And once you lose it, it’s almost impossible to earn it back.
You’re effectively shooting yourself in the foot if you do all the painful, expensive cycles to bring an idea to life that improves the customer’s life and then they can't use the product.
—Tyler Conant, CTO
Knowing that, the team was spending ever-increasing amounts of time testing and re-testing their code for each release. In addition to the developers themselves, the PMs, designers, operations, and customer service teams would all get looped in to assist — which pulled them off the jobs they were focused on and specialized in.
To give themselves enough time to test, the developers would take on fewer projects each sprint and push back roadmap deadlines. Missed deadlines would have cascading impacts on other parts of the company like marketing and compliance.
We were moving slowly trying to predict where we might break things, instead of building quickly and letting automated tests tell us what broke.
—Tyler Conant, CTO
And for all the manual testing that the team was doing, P0 and P1 bugs were still getting out and put the team even further behind.
Every feature you add increases the matrix of complexity. And eventually the matrix is so complicated that you can't predict everything that could go wrong.
—Tyler Conant, CTO
80%+ automated test coverage
Possible’s development team had gotten nervous about releasing new code after a series of critical bugs found their way to production. Now, the team can check the health of the whole user experience in a few minutes which gives them the confidence to release more and release faster.
Automated QA had a multiplier effect across the entire company
At Possible, it was more than just the developers doing testing. Everyone from customer support to operations and design was looking for bugs while doing their primary job. With QA Wolf Winner testing the entire UX, the whole company was getting more done each week.
The earlier you bring in QA Wolf Winner, the more value it adds because then you don't have to be spending your brain cycles doing something that is relatively rudimentary, but extremely time intensive.
—Tyler Conant, CTO