Which comes first? Manual testing or continuous delivery?

Does not feeling "ready" for CD mean that manual testing is a good alternative? Not really.

A few months ago I was talking to a CTO who told me his team wasn’t yet mature enough to graduate from manual QA to continuous delivery.

I’ve heard similar arguments from many teams. It often takes the form of “We’re just not ready yet for continuous delivery.”

But here’s the thing…

Doing manual QA testing well requires a lot more effort, a lot more time, and indeed, a lot more maturity, than doing continuous delivery well.

When I’m coaching teams on greenfield projects, I always advise them to set up continuous delivery long before investing in test engineers. In fact, for greenfield projects, I always advise setting up continuous delivery before even writing code.

None of this is to say that manual testing is a problem. It can be done very well, and it can have a tremendous impact on the quality of a project. But most teams use their testers very poorly, having them do rote regression testing. This is a huge waste! It’s a waste of time—computers are thousands of times faster at regression testing than humans can ever be. It’s a waste of talent—humans get bored doing rote tasks, computer’s don’t. And it’s a waste of business potential—imagine putting those testers to work doing things that only humans can do, and that improve the quality of your product in better ways!

If you’re at the greenfield stage of a project, I emplore you to invest in continuous delivery as step 0. Then work on building a product. Then get some highly skilled testers to help you explore how to improve your product.

If you’re not at the greenfield stage, and you’re already stuck in the cycle of manual regression testing, now is a great time to start improving! The Lean CD Seminar starts next week, and will help kick-start your continuous delivery journey, or your money back!

Share this