Why should failing in production be an integral part of your quality strategy?

In case your reaction to the title of this blog is somewhat closer to the above gif or any of the 22,800,000 gifs that show up when you search for ‘shocked reaction gif’ in google, let it be known that it is a perfectly sane and acceptable way to express the inner workings of your heart. In fact, I would probably be worried if it isn’t remotely like this.

‘Why in the name of ‘iFrames’ do I want to have something fail in production?’, you may very well hear this question echo from the depths of the left hemisphere from inside your cranium. Well, please relay this back that it’s not about ‘want’ but rather about the word ‘can’. Nobody wants to fail in production, not even the testers (in case you thought we did).

To even fathom that this idea of allowing failures in production could be good for the business, one might have to take a deep breath and begin a journey of introspection. And this journey of introspection is always laden with some hard questions, conceiving inconceivable scenarios, hypothesizing things that are beyond the comprehension of a brain that is only about 200,000 years in the making, etc.

‘Now that I have read 3 paragraphs, I still don’t see your point coming up on the horizon?’, you may ask, except of course if you’re a flat-earther and in which case this question may render new heights of ‘paradoxism’.

Just humor me.

As the first step in this journey of enlightenment, can we all agree that when the organization needs to cut costs, the first place the CIO goes around looking for budget cuts is the testing unit? Even better if it’s the test automation team?

Could it be because the CIOs across the world have an inherent phobia of the testing unit? Of course not (well there could be a few!). It is most often the case due to the amount of risk that the CIOs may be willing to accept. The ‘appetite for risk’ almost always determines that operational excellence begins with the testing team.

Let me introduce our old friend ‘Risk Appetite’. The ‘Risk Appetite’, in layman terms, and in this specific context, is the business accepting the extent to which the application can have defects in production. ‘This is madness! Why would the business want to have a product feature fail in production?”, I can hear the mind voice of some of you puritans. Mind you, what I’m suggesting isn’t a blanket rate on ‘can have’ production defects but rather a well-thought-out strategy to deprioritize features where the occurrence of defects isn’t going to hurt the business much.

‘Why? Just why?”

Simply because this enables us to

a) move less impactful features to the end-users for them to navigate, test and provide firsthand feedback (and determine if they are defects at all)

b) ensure the testing team can focus on extensive testing of the features that are critical to running the business rather than prioritizing few test cases across all features high and low

c) have a real feedback loop with the end users that can help show a mirror to the business on if they are indeed focussing on the right things

This task of determining what can fail in production may not be as simple as I make it here but definitely not as big as launching a rocket as well. Each business has its set of critical features that add value to consumers. The key is to identify those and well.. just leave them alone. Let's take the example of Instagram. When they made changes to the layout, moving around the buttons and icons, that did not exactly turn away the users although the user experience was impacted and the muscle memory had to be retrained to get back to browsing on autopilot. This was because the users find their value in sharing their images, getting instant feedback, and browsing relevant content that helps them in their own way. Rest is all noise.

Similarly, each business absolutely has to know what features bring home the dough for them and focus all attention on these features. The lesser impactful features can be left for the end-users themselves to experience and test.

This is not a new concept by any means and is an everyday affair in successful start-ups across the world. Their focus is always on the core features and there is no reason why enterprises, SMEs shouldn’t do that as well. The critical part however is how you complete the feedback loop from the end-users. Be it a feedback form in the product, or displaying a customer support email, phone number, or even using tools like ‘unbounce’ to measure user behavior on the site, the business should find ways to get that user feedback.

Let me pause here for a minute. If at this moment, you still think this is just about testing and the cost-cutting alone, I suggest you take a deep breath, have a glass of water, and re-read from the beginning.

Lastly, let me leave you with a question to ponder. If there are mismatches to the defined requirements in production and the users are perfectly fine with the way it works, would that still be termed as a defect?

Related posts:

Adding backing vocals to your recordings, whether at home or in the studio is an essential part of the recording process. It adds weight to your overall sound and really makes your lead vocals shine…

You may have been trying for months. Years. Or it may have happened quickly. You may have thought it would never happen. You may have not been expecting it at all. But one day, there it was. And you…

Like millions of people around the world, I visited Disneyland for the first time as an impressionable 9-year old and fell head over Minnie Mouse ears in love. The Enchanted Tiki room and the…