Testing and the rest of IT have a problem

The latest issue of the World Quality Report shows that more and more organizations and teams are struggling to get the right level of focus on all the rapid changes we are facing. The realities of the speed of change driven by Digital and the shift in IT functions towards Agile and DevOps have exposed the fact that people are falling behind the pace of change. Even when IT moves to the faster delivery model, the rest of the organization lags behind or doesn’t want to shift its way of doing business (there’s a term for companies that don’t adapt : bankrupt) Here at Sogeti, we don’t see this pace of change slowing down anytime soon. In fact, there is a very good chance that it will increase even further as more IT functions are driven to adopt continuous delivery to support their business needs. If your client takes two weeks, then it’s going to be a brave (and probably soon unemployed delivery manager) that remains committed to a six months release cycle. With IoT and ever more connected devices adding to the information and the complexity, how is the IT function going to cope? Automation is going to be key. But not functional automation. It’s got to be automation across the entire life cycle, everything that can be re-used needs to be re-used. Everything that can be done faster by a machine needs to be done that way. Development and Operations need to embrace the automation. Testers should be writing a script one and then the automation takes over. Testing will be wrapped around the ideas and shift to the left even more than we have already seen. More automation, more points of connection to the systems will mean more data. A lot more data. More than we humans can process in a reasonable amount of time. There will be more patterns that need to be explored and examined. Traditionally we have had to report on what happened. Perhaps a day ago and by the time we had a meeting we had gone several days. In a two weeks release cycle, that means your information is very out of date. In a one week cycle, it’s worse. But if you are delivering every day. Or every hour or less. Then that data is so out of data that it isn’t meaningful and worse, is likely to lead to out of date being used to make the one decisions, possibly with results that are BAD™ So what are we going to do? The answer is in the code. Machine learning, AI, call it what you want. The volume of data increases at a frightening rate. But data is useless if we cannot make sense of it. We need to reach the information that is in the data and human brains are very good at seeing patterns. Unfortunately, they will see a lot of patterns where there are none and confuse correlation with causation. We need help. A lot of it. And that’s going to come from the next generation of intelligent machines to help us in the increasingly fast changing and fast delivering world. Looking at what we have done, where we are now and where we want to be. The winners will be those who get a quality product out ahead of their rivals. It doesn’t have to be the best. But it has to work. Get to market sooner, with few defects means a better revenue stream, with a lower risk profile, and you don’t have to be best. Just don’t give people a reason to change.