Manually Executed Software Testing:
Out of Fashion?
Do you (still) enjoy software testing but do not want to be a developer in test?
Have you seen all the job adverts for testers and decided there's no future for you without learning some programming?
I hear from plenty that feel like this.
You're probably, and necessarily, fine with running some debugging or querying tools and small utilities to add some repeatable automation to your testing.
Have you seen all the Agile project roles advertised and not seen a progression for you into test management?
The software testing world has changed and if yours hasn't, it likely will.
This article was intended as a 'comment' to my recent LinkedIn post about hiring manual or automation 'testers', but it was too long for the Comments field :-) The original post was:
Imagine, you have budget for a small test team.
Do you hire 1 manual tester, a crackingly good one, and 2 automation testers?
Or, do you hire 2 top manual testers and 1 automation tester?
I have been astounded by the interest in this post and the healthy debate and well argued opinions in the comments.
I anticipated maybe a couple of hundred views and 5 to 10 comments. It's now nearing sixty thousand views and 85 comments!
It obviously touches on a (or more than one) poignant topic at the moment.
Thank you all for your comments on the post.
The question posed was almost word for word what I heard when talking to a small technology company.
They'd seen quality issues in their 'Live' product and needed a more structured approach to testing.
They saw a pipeline of much increased sales and decided 'to take no risks' and 'sort our testing out'!
They'd heard from their programmers that they only needed some 'automated tests' but I advised some exploratory 'manual' testing first before trying to automate tests.
So, that's the backdrop.
Yes, there have been a few different inferences taken from the sparse context my question implied.
I expected that and worded the question to allow freedom of ideas based on individual thoughts.
I didn't want to lead anyone towards my thinking just to have them confirm it in my narrow context.
There is never a 'right' answer but there are several sensible approaches based on the sketchy detail in the question.
Yes, there has been evidence of differences in the understanding of terms/concepts.
Again, I deliberately wrote the post to allow this to happen as sometimes we talk the same language but mean different things; or we use different language but mean the same things.
From this post, most importantly, I wanted to see what value people place on humans testing, these days.
My feeling was that many people believe they can eradicate manual/exploratory testing in favour of some automagic testing. This has been true for many years but rather than becoming less widely held, I feel the belief is spreading with the increase of SDET roles and DevOps and Agile dominating. The drive for speedy product release seems to make us turn to automation.
Most responses to my post suggested hiring one or more 'automation tester'.
Some suggested only manual/exploratory testing, with developers programming automated checks at some reasonable point.
Other opinons were somewhere in the middle.
It has shown me several things:
- Plenty of people (testers and programmers) still value the human and exploratory testing activity and would rather trust this to mitigate risks they see in a product.
- Some people still believe automating all their tests is the way to go :-(
- Many people's comments are a product of their current or most recent work experience in a particular development lifecycle/methodology and do not always take into account other working landscapes out there.
- Some people stick to the supplied answers while others apply much more thinking and question the question :-) This I liked, particularly.
- Despite many advocates for skilled exploratory human testers, job adverts are very much weighted against this.
As with other clients in the past, I will continue to advise, in similar scenarios, exploratory testing to begin with and then phase in some developer created automated 'checks'. Exploratory testing should continue, whilst expanding the automated suites, and beyond.
Automagic does not exist, only in the heads of those that have not automated any checks/tests.