Do we need Software Testing and Test Automation ROI (Return over investment) calculations?

By on March 6, 2013 Comments

So what would be the ROI of reading this blogpost? It’s possible that you get an idea which helps you in your test automation effort or removes myths you may have about software testing or you find that you are in software testing field because of wrong reasons and you leave this field altogether.

There are many possibilities and all of them are difficult to quantify. In my opinion, it is difficult to quantify most of the activities we perform in our field - software testing.

I have never done ROI calculations for any testing activity. My take on most of the activities I perform is simple and based on the MoSCoW model

Must

Things like exploratory testing, robust test automation, Continuous Integration, NFRs etc are more or less MUST for projects I get involved in. We discuss and get agreement on how to to implement / manage them but never on whether these should be done or not.

Should

Things like screenshots on failure, randomization, cross-browser testing, increased coverage, testing dashboard across project, visualization etc are kind of activities I tend to push, if there is any bandwidth. However, I never use arguments like - if you use defect tracking system 10 times in a week, it will take X min and cost you Y GBP based on your hourly rates. However, if we extract, visualize and put this data in on a dashboard, it will take 2 days and will save you Z GBP over a period of six months.

In my opinion, cost is an important factor in decision making - but it should not be the only thing you have when you sell your idea. If there are no other benefits of extracting this information and putting it on a dashboard except saving Z GBP over a period of six months - may be it's not a good enough idea for should category - who knows if usage pattern will remain as it is for next six months or not?

Could

If we run out of Musts and Shoulds, we can look at things like Archiving results, sessions, reporting around exploratory testing sessions, increased coverage and depth of testing, development of tools / scripts to help us in the long run and so on..

Won’t

Oh, there are plenty of things in this category and I do not need ROI calculations to find that they are wasteful :-)

My decision making (MoSCoW model in mind) is pretty simple - it is based on either my gut feeling (or a pattern I might have seen in past) or it is based on the opinion of someone I respect (who have demonstrated their expertise in the current project) or it’s decided by the team after weighing available options.

I have never performed any ROI calculations to demonstrate the need of testing, automation, CI, buildpipeline etc. I have often used discussions & reasons to convince whenever needed.

I have mostly worked in an environment where ROI was never requested / needed - but what about service industry or outsourcing. Do they use ROI a lot, is it justified? What do you think about it? Have you done any sort of ROI calculations to demonstrate the need of testing, automation or something else? Did you find it useful? Please leave your comments and we can discuss.

blog comments powered by Disqus
Finished reading? Browse all posts »