Wednesday, November 18, 2015

The success rate percentage predictors of training a manual QA person to become an automated tester

Since 2010, every job I have had has included one task that is particularly difficult to achieve: train and convert a mostly manual QA tester (or team) to be an automated tester.

Most timeframes include 3-6 months to accomplish this task.

I have never had a 100% conversion rate. I have never had a 50% conversion rate. The average is most likely hovering around 30-40%.


Well, several factors determine how successful a manual QA person transitions to be an automated tester. And by knowing little bit of data beforehand, I can almost accurately predict which people can successfully learn test automation skills and write excellent test automation to augment their manual tests:

1) Software education background and coding experience - this one is the most relevant in predicting the outcome of the QA person making the transition. This contributes 60% to successful chance of the conversion.

2) Familiarity with one or more of the testing technologies chosen for implementation  - for example, SQAEvangelist uses a stack including Ant, JUnit, TestNG, Selenium/WebDriver, and Java. This contributes 15% successful chance of conversion.

3) Ambition - believe it or not, this attribute contributes a lot. 15%.

4) Work ethic - good intentions and great work habits don't help to succeed much - 5%.

5) Training - someone spend 1:1 time and helps the tester understand how to do it and even goes through examples. 5%.

Remember that these numbers do not guarantee failure or success or each person making the transition. It's a hard and impossible road for some QA people.

People with attributes 1 and 2 are nearly a lock to make quick contributions to the automated testing effort with a few days.

I have found that any person with attributes 1 and 3 are almost certain to become a good automated tester and can almost be efficient and make significant contributions.

So before you challenge the team to learn automated testing, consider these factors before making a final decision. Analyze the team and predict it's chances of success beforehand.

It may be easier to hire/create an automated testing team to automate and just let the manual testers stay manual testers.


  1. So, if I get you right, are you saying that the probability of learning to program at work is very low, and most of the time trying to teach someone programming will be a waste of effort?

  2. Oh no, not at all, of course not. How did you come to that conclusion from this post?

    I don't think the effort by any person to learn something new is a waste of effort.

    I spend a lot of time mentioning the teams and some team members will be able to do test automation for several reasons, and some won't be able to. The post was to indicate which factors matter most.