I had the pleasure of sharing breakfast with Gojco (pronounced “Goyco”) Adzic at last week’s STAREAST conference. Adzic delivered one of the keynotes at the conference, title “Sleeping with the Enemy. He talked of the need to change the traditional culture, in which testers work independently from developers, to one in which there is trust and teamwork between business, development and test.
In fact, in the short video clip I took, Adzic listed culture change as the most difficult problem that testers face. Naomi Karten addressed the challenge of organizational change in her keynote as well, talking about the chaos we all face during transition and the need to give people time to adjust to that change.
One cultural change that Adzic and I discussed was around how we measure quality. We were chatting about Lisa Crispin’s upcoming presentation about defect tracking and I asked Adzic whether or not he felt it was possible for managers to get the metrics they needed to make decisions if a defect tracking tool wasn’t used.
Adzic answered rather emphatically that the number of defects is not a good measure of quality. He reiterated some of the lessons I’d heard from Lloyd Roden’s session about challenges testers face, in particular the misleading nature of metrics.
Adzic felt that rather than using defect numbers as a measure of quality, managers should be looking at things like customer satisfaction and product revenue. What good is a product with no bugs if no one wants to use it?
He has a good point, but can QA managers make decisions or recommendations about the readiness of a product to go live without gathering metrics on defects? It’s an interesting question with differing viewpoints. What do you think? Is this another area where our culture needs to change?