User testing overview

This video looks at how user testing can provide a more complete understanding of the customer journey, and why people drop out of your sales funnel. It looks at who, and how many people, to use as testers, along with how to set up remote testing.

To find out more about, please use my affiliate link:

If you'd like to read more about why you only need to test with 3-5 users, Jakob Nielsen's article can be read here:

Click here to hide or show the transcript


One of the key challenges with improving your site’s user experience is the bias and knowledge you bring to the task. Even if you take an objective look at your site, because you know how it works, it’s really hard to see it with a fresh pair of eyes.

This is where user testing can really help. In this video I’ll talk you through the benefits of user testing and how to set up your testing to get the most out of it.

You can use your actual customers as the testers, by inviting them into your business and recording them, or you can use an agency to provide you with a target audience like your customers via demographics and screening questions, who complete the testing remotely, in their own homes. Or, you can use a mixture of the two - an agency to supply the recording software and analysis tools, but send your own customers a link to become the testers.

During a user test, the tester visits your site, whilst their screen and audio commentary about the experience is recorded. This allows you to really see where the friction points are in their journey, and will usually also supply feedback about their expectations and where your site fell short. If you have a point in the transaction process where you’re seeing a lot of visitors abandon their cart, or if you have a section of your website that you expected to be popular but which isn’t getting much traffic, user testing can help you to find out exactly what’s happening. Even if you’re not aware of any problems, running testing can highlight areas where the journey could be smoother - you may think your site is working well because you’re getting conversions, but maybe those conversions could happen more quickly and easily. It can sometimes be hard when using Analytics to determine if metrics like time spent on page show engagement or confusion. User testing gives you that insight.

It’s vital to select the right type of person to test. They should be as close as possible to your core customers - so if your primary customer is men in their 30’s who listen to music on vinyl, aim to test with that exact demographic. If you’re using your own customers try not to include only long-standing customers - as well as the issue of their greater loyalty making them less willing to criticise, if they are repeat visitors to your website they too may have learned some of its quirks and be less aware of them.

You only need 3-5 people to test. Jakob Nielsen, of the Nielsen Norman Group, has found that 5 people will find 85% of usability issues, and more than this isn’t an effective use of your budget - a larger budget should be spent on more small tests.

Unless your site experience is vastly different on mobile, tablet and desktop, I would advise testing on the device the majority of your visitors use first, make any revisions, and then only repeat the process on a second device if conversions don’t improve amongst visitors using that device - if you have a responsive site, it’s likely that the majority of usability issues are with the general flow of the journey rather than how your page looks on different devices.

The structure of the test is also very important. When you plan to run user testing, it’s not as simple as just putting someone in front of a laptop and asking them to visit your site and comment. With such an open brief, you’re unlikely to get much insight, and if you’re working directly with your customers, they may be reluctant to criticise your site, seeing it as impolite or confrontational, or they may take the opportunity to tell you something that annoys them but which isn’t what you’re looking to get out of the exercise - for example that you no longer stock something they used to buy.

If you’re using a sample audience, it’s good to start with an opening task of “When the site has loaded, say what your initial impressions are, who you think the site is for, and what you think you can do on the site.” This can help to highlight any general misperceptions, which may influence their later behaviour, and can give some good general insight, as well as easing the user in with a quick and simple exercise to allow them to settle into the testing.

Once the user has given their initial impressions, the best way to get something of value is to give them tasks to complete which relate to what you want to test, and ask them to comment on their experience as they complete the task. In this way you’re not asking them for a subjective opinion, you’re seeing how easy it is to navigate and use your site, with a black and white, success or failure outcome for completing the task. It’s also good to give users a scenario for the task. This helps them contextualise it and makes their commentary and actions more natural. Obviously the nature of the task will depend on what aspect of your site you’re investigating.

As an example, a furniture retailer is finding they’re getting a lot of emails and calls asking for information about product dimensions. So they want to check whether users are navigating products easily and then finding the product specs, which they have to click on a tab to view - the default view shows the description. They might set a task such as “Imagine you are looking for a new sofa for your living room. Your current sofa is a two-seater and you want to upgrade to a three-seater. The sofa needs to be 210cm long or less to fit your room layout. Browse the sofas and select a suitable one. Please comment on what you are doing and why as you complete this task.” The user will then need to navigate to sofas and find three-seater models, and find the details to check the dimensions.

The test could then go on to ask them to “Continue to purchase your selection, stopping at the point of you need to enter credit card details.” This is a common task, as it will allow you to see how easy it is to complete the checkout, including any delivery options. If you have dummy card details, you can also give these to the user to enable you to see the whole process - this can be useful if you are finding people are placing duplicate orders or you’re getting calls asking you to check if an order has gone through. You should obviously never ask users to enter any real credit card details or details of a personal or protected nature. You’ll also want to use dummy details for any form submissions you ask them to complete - not only does this protect the tester but you can then identify the false submissions and remove them before they’re processed.

Some user tests also ask users to comment on how easy they found it to complete the task - the time taken and number of clicks will probably give you a good indication without this step, but it can be good to add some short questions like this to keep the user vocalising their thoughts.

Ideally tests should last 15-20 minutes. This will:
• enable you to give the users a few tasks to complete
• ensure users don’t lose focus or rush tasks
• provide a manageable amount of data for you to review

If you use an agency for the testing, they will pay the testers for their time. If you’re using your own customers you may wish to incentivise them to complete the test by offering them a discount code for remote testing, or if you’re completing the testing on site you could offer them free refreshments or a low value gift. I would advise setting the user up at a computer in a quiet space and leaving them too it - they will feel more comfortable without someone hovering and you’re trying to replicate a true browsing experience, which won’t happen if they feel they can ask you for ‘help’. Try and put them at their ease, explain that you’re testing the site, not their ability, and ask them to be as open and honest as they can and vocalise all their actions and thought processes.

If you do choose to pay an agency, don’t be afraid to reject unsuitable videos - if the sound or visual quality is poor ask for a new tester. Also, if you receive a video which is over twice the length of time it should take, I would reject this too - it’s likely the tester is having issues beyond those of your website! Plus, your time is precious too and if you’ve set a 15-20 minute task you want to spend only that amount of time reviewing it.

When you review the videos, you can make a note of issues or friction points, and you should start to see some common ones amongst the users. If you’re using an agency, you may be able to annotate the videos as you watch them, within the software they provide. You can then export all the annotations for review. Your findings will then form the basis for your future web development and should result in less friction and more conversions as a result.

Thanks for watching this video. If you’d like to begin user testing using an agency, check out the following links:

I have used WhatUsersDo in the past and they have brilliant customer service and will really work with you to get results - you get an onboarding session to help you plan your first tests and review the data, with follow up calls and lots of account management and hand holding. However their pricing seems to have increased quite a bit since I last used them - it’s currently £650 for 10 tests is pretty high. They offer a free trial, which will give you an idea of the service and how their system works. was set up as a challenger to the bigger organisations with a determination to make user testing more affordable. I haven’t used them personally, and they currently only provide desktop testing, but they do have an account manager. At the time of making this video they’re charging £18 per user, so if you believe desktop is representative of your mobile and tablet experiences, I would recommend giving them a go. If you use the affiliate link below, you’ll get one free tester. They also currently offer a discounted price of £8 per tester for your first test.

This link is an affiliate link and if you use it I will also receive a free tester, however I’m recommending from my research into alternatives to WhatUsersDo, speaking to their representatives and a genuine desire to promote a product that seems to provide great value. If you wish to purchase without me receiving any benefit, just run a search for their name.

If you'd like to read more about why you only need to test with 3-5 users, Jakob Nielsen's article can be read here:


0 Comment

Leave A Comment