Automation - The Purpose
I Read a lot, and I am active in quite a few places on the internet. This post came about from some free-form journaling based on a post about test automation from LinkedIn.
The post asks some questions about the link between automation and defect finding. It makes statements about automation the general subject from a specialist test-automation context.
Where I am at.
There are several cases for automation, and while defect finding and reduction is one, if it is your foundational case, you will be so wrapped up before you launch anything; or so falsely confident in your ignorance, you will lack benefits of automation anyway.
I Believe that by completing automation of even a defective state first, it should reflect the reality of where you are at and have been.
I have great respect for the past. If you don't know where you've come from, you don't know where you're going. I have respect for the past, but I'm a person of the moment. I'm here, and I do my best to be completely centered at the place I'm at, then I go forward to the next place. Maya Angelou.
Quick aside on improvements.
As with a lot of things, by committing a process honestly to a record, you will find improvements. Resist the urge to make changes before you deliver the honest record of where you are currently at! Squirrel away notes on how to improve, thoughts, things to come back to later; but do not spend too much time on them before finishing.
A Step-by-step guide.
Document your processes.
This is the primary benefit of automation hands-down. There is not one single benefit that outweighs being able to honestly understand what you do. If you do not have a (hopefully granular) breakdown of what you do, and how it is done. Your risk of failure is multiplied, you will struggle in communicating your value and what you do.
Monitor consistency and enforce standards.
If you wing-it, every time you do something, then your risk goes through the roof. If you are not consistent, then you are winging it. Following the documented processes, knowing where you deviate can also help you to iterate on that.
Having a process is not enough! This might mean instrumenting processes with checks and logs, creating incident logs, template checks, manual QC (not QA yet). QA is not something you can bootstrap.
Cost reduction.
If you never reduce your costs, and it costs you the same $50; to get something done as it would with a person for the time, then you will have to conjure magic money from margins or elsewhere, in order to be able to action or identify defects or make any changes.
If you have not got points 1 or 2 done, you cannot effectively participate in this activity, and without it you cannot effectively do anything else consistently.
If you did manage to find "magic money" for this ahead of the pre-requisite steps, you would be iterating blind, and will likely at best get waterfall repetition.
Identify defects and iterate.
Once you have a process, and a framework, and it has realized cost reduction, you can make mature, intelligent decisions about where you are going knowing and being able to eschew the benefits and choices behind your decisions.
Defects are in my mind at least different from cost reductions. A Cost reduction might mean you use aluminium rather than steel. This has properties, that might showcase defects in decision making, or vendor selection. It is not the material, but the process that is problematic.
Wrapping up.
This is all very high-brow stuff. Many people will look at this, they may scoff or continue to operate as they are. My views are the product of nearly two decades working for numerous companies and organizations around the world, seeing what I felt worked, it is the closest to a controlled experiment I have, but they might still not be right for you and are not yet verified facts.
Try out the 4-step guide for something small, but strategically impactful. Let me know by posting a response using webmentions if you find improvements, or have other experience, evidence or facts.
Very few people get this far, and I cannot claim that I universally achieve completing this 4-step cycle.
If you skip ahead, as I have many times. I Promise it will come back to bite you. It always bites me. Also, if you follow the process, you might identify early on that something has to go, so you will hopefully be able to channel your efforts.
- Complete automating your current process or bin it.
- Schedule future work via notes and journaling.
- Let me know how it goes.
I Wish you well on your automation journey and hope you learned something or enjoyed reading.