Why manual web testing is irreplaceable

Posted by Dan Berry on Monday 16 October 2017
Find me on:

Web testing is a complex undertaking. There are many different testing services available to see how your site or app will perform. For example, you could test from your user’s perspective (usability testing) or check it works on every conceivable device, screen size and OS variation (functional testing).

You can also carry out either manual or automated web testing on each of these testing services, depending on your project requirements.

Automation testing (software testing software) repeats tests that are made up of predefined actions. In other words, your user does X and an automation test checks that Y happens every time. Such tests use an automation tool and are gaining in popularity with 56% of large businesses now conducting automation testing.

Manual tests are conducted by a human, who checks the functionality of the software much like a real user would. Again, the test may want to check that if our human user does X then Y happens - but now there’s a real person conducting the test.

So, which one would you choose? Regardless of your answer, I’m afraid you’re wrong. Because you often you need both forms of web testing to get the best results for your software.

Let’s look at the infamous Millennium Bridge debacle. It’s a perfect example of how you can’t fully test each and every scenario without some human interaction to highlight and uncover unexpected behaviours in your system.

When the Millennium Bridge was unveiled it wobbled more than a jelly on a gyroscope. It was assumed this was due to footfall, even though the bridge had been tested for large amounts of footfall. However, (unbeknown to the bridge’s designers) people synchronised their steps on the bridge, causing it to sway uncontrollably.

This outcome was completely unexpected. The synchronisation of steps and its catastrophic effects were not modelled when designing the bridge.

The Millennium Bridge is somewhat of an analogy for automation and manual testing. You can’t model for every nuance of human behaviour in your automation testing.

So, you need to adopt a mix of both manual and automation web testing to provide the best testing harmony and coverage for your software.

Web Testing black and white.jpg

But this (could) add further complexity to your web testing strategy. If you need a machine and a human tester to check every aspect of your website, then you may see spiralling costs and lengthy test projects.

So, you need to optimise your approach. You need to take the best bits of manual testing and automation to give each specific project an optimised series of tests.

What’s the solution?

You could argue that everything can either be tested manually or automated with the right tools and the right environment. However, creating these tools and taking the time to test and verify the results will swallow more of your testing time and budget.

So, make sure you consider your time, your resources, and the size of your project when you develop your web testing strategy. You may find that one type of web testing better suits a particular project.

For example, if you are developing a new user interface with only a handful of test cases, then manual tests will help you to identify any areas of weakness and examine the user experience in the most cost effective manner. But you may also want to incorporate some simple automated tests to click every link on the new interface, for example, and identify bugs that could be missed due to browser blindness.

Combining both manual and automated tests is often the optimal option. You’ll cancel out the pitfalls of both to create the best result for your website.

Using both forms of testing will ensure that you capture the human insight whilst automating the repetitive tasks. It’s the best of both worlds.

Need to revolutionise your web testing strategy? Download our guide.

Web Testing