Act natural
One of the biggest challenges for producing quality results in user testing is ensuring the tester behaves naturally, i.e. in the way that they would when using your product for real. Here's some thoughts on how you might get as close to that as possible.
For the sake of this post, I'm just going to talk about unmoderated/video user testing, as there are quite a few more variables in lab testing that I'll cover in another post one day.
Here's some thoughts about how you might achieve this when all you can control is the test itself...
1. Trigger their imagination
It's not really enough to say 'imagine you've decided to buy a new t-shirt', as this doesn't really generate all the peripheral thoughts and emotions that would normally go along with wanting a new t-shirt.
Add some more context and allow the user to build their own story around it, try these:
- "Imagine the last time you bought a new t-shirt, was it for a holiday? An event? Or just to replace old ones?"
- "What do you look for in a t-shirt? What do you like and dislike?"
- "How much would you really spend on a t-shirt?"
Even if some of these things are irrelevant to your test, it triggers the user's own imagination so the perspective they are using to approach the task is their own, not yours.
1. Don't ask them to do the thing
If your testers are reading and then following your instructions to complete a task, they aren't testing the thing, they're testing your instructions, and definitely not behaving naturally.
So the first thing to remember is, if possible, do not ask them directly to perform the task in question; try to put them into a scenario where they might decide perform the task themselves. This way, they are more likely to be behaving naturally at the point you need them to be.
3. Discourage 'Tester mentality'
For the most part, testers actually engage very well with whatever system you put in front of them. Some, however, look more subjectively at the situation and try to apply more thinking than instinct to it.
This 'tester mentality', is a real problem. It is typically seen as a tester trying to pull apart the design to figure out what they're 'supposed' to do and comment on whether they think the design is likely to achieve that for most people or not.
In my experience, the only way to reduce the likelihood of this happening (since you can't correct it in unmoderated testing) is to give them something that engages them further and causes the tester to have more invested in the experience. For instance:
Give them a voucher code to use to buy something; whilst this may create an unrealistic situation in some respects, they may at least be more concerned about finding the right thing to make the most of the freebie. Great for testing search and configuration features.
Ask the tester to do something quickly. Again, this may be problematic for some tests, but it can make it more likely that they will act on instinct rather than calculating their actions.
Distract them with false tasks. If you want to test a search filter for instance, be very specific about asking them to test something else that is likely to involve the search filter, such as the using the 'favourites' button to bookmark their favourite three t-shirts. Labour the point about asking them what they think of the favourites button and they'll focus all of their 'wisdom' on that.
It rarely gets as bad as needing to take the last measures listed above; user testing is a generally a very effective process for finding UX issues, but no one wants to waste time analysing test that aren't valuable so it's good to write your tests in such a way that you will avoid most of the common problems.