Upstreaming Quality

Penny Wyatt, QA manager at Atlassian, gave an insightful talk at the 2014 Atlassian Summit on how to do QA.

QA at Atlassian, and other companies such as Spotify, now stands not for Quality Assurance, but for Quality Assistance.  Which makes perfect sense; a QA group in a modern, fast moving software company, is in no position to ASSURE quality ( and meet deadlines). But they have  excellent training & experience to help others to improve quality. Quality is everyone’s business, and quality needs help and the further upstream you go, the more quality assistance leverage you can get.

Its a natural thing, when   a  fast growing software company starts to get complaints from customers, to go out and hire a Quality Assurance manager and lots of Testers  to catch all those bugs before customers discover them. This seems to solve the quality problem.

But a second problem arises – finding bugs by testing is too SLOW.

The quality problem is further upstream  – those bugs  should never have been inserted to start with.

Upstreaming  thinking leads to making  quality improvements closer to the source, to developers, to analysts creating specs. For example, developers insert automation to testing to prevent regressions, they develop ways to reduce complexity and dependencies. Developers can be helped to learn to test, and think like testers, close to the moment a bug is created. Code reviews.  Detected a bug in a Requirement at requirement-creation, can save person weeks of coding and debugging and testing time.

The solution to quality is not more testing, or testers, its less bugs.