At the TestBash in Cambridge, Ben Wirtz talked about a tale of a startup, or when the chance of passing a test successfully is less than 50%.
Ben explained that he doesn’t know a lot about testing, but about startups. He wanted to talk about the value that testers can bring to startups. Ben developed an Android application (look for “Handy Elephant” (please don’t test it, he said)). He did the bootstrapping for a year. He eventually received some investment. They changed their business model in the past – a lot. Despite the changes in business plans, they tried to reduce the technical risk – at least that was the plan. It eventually turned out that the changes in business plans eventually leads to a sinking ship with too much technical debt incurred.
The chance of anything passing a test successfully is less than 50%. Ever, Wirtz explained. From day one startups are jeopardizing to become a sinking ship. Lean Startup therefore focusses on learning. Experiment (and deploy) continuously, and pivot your strategy if you need to. This refers Build-Measure-Learn loop that I also found in the Lean Startup book from Eric Ries. Through testing hypotheses you accelerate learning. By automating a lot of tests you can make changes in your business model while still keeping your tech working through practices like Continuous Integration and Continuous Deployment. The main objective for a Lean Startup (or any Startup for that matter) is to find the value proposition and your business model through accelerated learning.
Ben explained that they hadn’t done much manual testing in their environment. Instead they work from a highly automated testing basis at the code level, so that the startup can pivot quickly based on this. The advantage of being a startup is that users – Early Adopters – are often forgiving. Users are also often enthusiastic about it.
Traditional testing and quality assurance is usually mostly about not screwing up your product. In a startup this kind of manual testing is not that helpful. In a Startup the situation is shifted. You are not too sure about your business model, and will find out more and more about it on your way. Therefore building the perfect app is not a goal of a startup. That’s why more traditional approaches to testing likely fail in Startups, I want to add.
As tester raise constructive criticism with questions like “what does the user actually want to do?” or suggesting alternative implementations. He also added that if nobody listens to you, focus on learning yourself, and start your own company. Ben reminded us that we should assume that the app will fail also even after passing all of the automated tests.