preload

Alternative Facts? “QA Is Too Expensive”

Posted by Albert Gareev on Apr 26, 2017 | Categories: DiscussionsNotes

This very phrase “alternative facts” recently has become something. It is used mostly in a sarcastic way. With this post I’m starting a series on alternative facts about testing and quality assurance.

Let me begin with a couple of stories from my experiences.

Story One

Software development team works on a set of features for a new release. Programmers and testers review requirements; the code is being built and deployed; testers perform testing. Testers find a few bugs. Programmers pinpoint and fix them; deploy the updates. Testers verify the fixes and check for regression.

On the project balance sheet it ends up as testers spent nearly twice as much time on the features than programmers. All hours are billed to the project account.

Business stakeholder reviews the numbers and makes a conclusion: QA is too expensive.

Story Two

Software development team works on a quick patch for production. Fixing a bug takes just one hour. Another hour is dedicated to verification of the fix and regression testing. Deployment scripts push code to production. Testers do post-promotion testing in production and discover that the fix is not there. Apparently, something went wrong with deployment scripts. Programmers repeat the procedure. Testers repeat the post-promotion testing.

On the project balance sheet it ends up as testers spent nearly three times more hours while working on the patch than programmers. All hours are billed to the project account.

Business stakeholder reviews the numbers and makes a conclusion: QA is too expensive.

******

I saw this happening for real. Saw more than once. Can we call these alternative facts?

I’d suggest that this is an alternative interpretation (mis-interpretation, really, but who likes to be told that they were wrong?).

In both situations, time tracking, the common book-keeping method prompted the wrong conclusions while providing right numbers. But the root cause is the way the organization thinks about testing. As long as you think of testing as an “overhead” to development costs any number of hours would be seen as too expensive.

Managers need to make decisions about software. Will the software product fulfill the claims? Are there any collateral damages? What are the risks? Testing can provide information that reduces the uncertainty and helps to mitigate risks.

In the both examples, the information obtained from testing helped to avoid incurring of costs. But those costs weren’t estimated and reflected on the balance sheets.


  • One response to "Alternative Facts? “QA Is Too Expensive”"

  • Jim Hazen
    26th April 2017 at 8:55

    Your last piece hit it on the head. Is Testing really more expensive than having an escape into Production. No, and because of the additive effect of going through the whole cycle again to get the Production level defect fixed versus the one in-process before release. It is always cheaper to find and fix a defect early on than later on. That is what is lost on some people who do not understand the whole process. Testing isn’t free, but it can keep both your hard dollars and soft dollars for a project from becoming out of control.

    [Albert’s reply. Thank you, Jim. I agree. Unfortunately, still so many people don’t understand development and testing process.]

Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported
This work by Albert Gareev is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported.