← Back to all posts

Part 3 of 6 · Manual Testing · 5 min read

The manual tester is not going anywhere

I spend a large part of my week building automation. So it may surprise you that I argue hard, in almost every planning meeting, for keeping strong manual testers on the team. Automation and manual testing are not rivals. They do different jobs, and a team that forgets the second one ships worse software. Here is the case for the manual tester, made by someone who loves automation.

What a machine cannot do

It cannot be surprised

An automated test checks what you told it to check. It confirms that the total equals the expected figure and moves on. A person notices that the total is correct but the currency symbol is wrong, the page took four seconds to load, the button sits half off the screen on a smaller laptop, and the confirmation message has a spelling mistake. That wandering attention, the ability to notice the thing nobody asked about, is the whole point of exploratory testing.

It cannot judge how something feels

Is this flow confusing? Is the error message rude? Would a first time user understand where to click next? Does this screen feel slow even though the numbers say it is fast? These questions decide whether people enjoy using a product, and only a human can answer them.

It cannot test what has no script yet

Every new feature arrives before its automation. Someone has to be the first person to really use it, to probe it, to break it, and to decide which behaviours are worth locking down in an automated check later. Manual testing is where that understanding is built.

It cannot follow a hunch

Experienced testers develop instinct. Something about a change smells risky, so they poke at a related area that was not in the plan, and they find a defect nobody expected. You cannot write that instinct into a script, because it comes from years of pattern recognition and deep knowledge of one specific product.

The work that keeps quality high

Good manual testers do far more than click screens. They review requirements while the ink is still wet and ask the awkward questions that save a sprint. They design the test scenarios that automation is later built from. They run the exploratory sessions that find the defects a regression suite was never written to catch. They handle the cases that are slow or costly to automate, such as complex device setups, printed output, or a rare third party integration. They sit with real users and watch how the product behaves in the wild.

The healthy balance

Automation should carry the repetitive, stable, high value regression checks. It runs on every commit and frees people from drudgery. Manual testing should carry exploration, judgement, first contact with new features, and everything that needs a human sense of what good looks like. Teams get into trouble when they treat automation as a replacement rather than a partner, chase a headline number for automated coverage, and let their exploratory skills fade. A year later they have a green pipeline and a product that feels wrong in ways no test was written to catch.

The bottom line

Automation tells you the software still does what it did yesterday. A skilled manual tester tells you whether it does the right thing, whether it feels right, and what is about to go wrong. I want both on my team, and I am not willing to trade one for the other.