Your Online Presence > Strategy & Development

Website user testing on a budget

By Melanie Burke

How usable is your website? Do your visitors find what they need? What kind of experience do they have? Will they come back again? You can answer these questions by carrying out user testing on your site and you can do it on a budget.

User testing involves getting together some typical members of your audience and asking them to carry out tasks on your website whilst you observe. Watching how they get on, and listening to what they say, will give you valuable insights as to how you can improve your site.

User testing has got a bad reputation in some circles as being expensive and time consuming. After all, don’t you need a lab, moderators and expensive equipment? Fortunately, times have changed and it is possible to carry out simple user testing on a budget. Follow the advice in this article and it should take no more than 3 days to complete.

Taking the plunge

So, you’ve taken the plunge and decided to do some testing. Where do you start and how do you keep those costs down?

You’ll need to:

• Decide what to test
• Recruit participants
• Create context and tasks
• Test and capture the results
• Analyse the results
• Start all over again!

Deciding what to test

Firstly you need to decide what it is you are going to test. To keep it simple you should focus on key issues, areas of the website where you anticipate problems or any complicated processes.

Example: homepage, the navigation, the sign up process

Recruiting participants

Now you need to recruit your test subjects. Ask yourself:

• What types of people use your site?
• What are their motivations?
• What particular needs might they have?

Use this information to build up profiles of your users.

Example: person responsible for corporate social responsibility in their company.

In an ideal world you would use an external recruiter to find your test subjects but as this is testing on a budget you will probably end up doing it yourself.  It is better to test with people who don’t perfectly match your user profiles than not to test at all.

To find people to take part ask around, send emails and above all, provide people with an incentive to get involved. This could be a nominal monetary payment, a prize draw, or simply the opportunity to help your organisation.

Don’t recruit too many subjects; you will need no more than 5 people to get useful results. If you manage to find more participants keep them on file for your next round of testing. 

The length of each test should be no more than 45 minutes; this means you can easily carry out all your testing in one day. Keeping tests short is more convenient for your testers and won’t give them time to get bored.

Creating context & tasks

The next thing to do is come up with scenarios that explain the context of why your testers are visiting your site. You can create any number of different scenarios. They can range from quite specific to very general.

Example:

You are responsible for your company’s corporate social responsibility plan and you are looking for a charity for your company to sponsor. You want to partner with an organisation that offers projects for your employees to get involved in.

You now need to come up with specific tasks around each scenario. Aim for around 10 tasks taking into account your 45 minute timescale.

Examples:

• Does this organisation offer corporate fundraising?
• What kinds of projects do they run?
• How do you contact them?

Testing and capturing the results

Expensive testing often involves using a lab, two way mirrors and lots of equipment but all you really need is a quiet room with a table, two chairs and a computer with internet connection. If you want to record the testing you’ll need some additional equipment. You can either use a digital camera or software like Silverback (for Apple Macs) or Camtasia Studio (for Microsoft Windows) that allow you to record the mouse movements and the person’s face.

If you can afford it then you will benefit from hiring a usability professional, if you can’t then it is better to do it yourself rather than not at all. The best case scenario would be to record the tests and have one person facilitating. If you aren’t going to record the tests it is best to have two people conducting the test, one as the facilitator and one to take notes.

Before you let loose on real, live subjects you should practice the test on a colleague or other willing (victim?) person. A practice run will draw your attention to any inconsistencies or other problems in the way you have presented your tasks or the test itself.

Before you start the test you should explain to your test subject what the test is about, making it very clear that you are testing the site and not them. You can also ask them some background questions which will give you some insight into their level of experience or any notions they have about the site. It is useful to ask the subject to ‘think-aloud’ throughout the test, this will help you in understanding any problems they have.

During the test you should read each of the scenarios and tasks to your test subject. It might help to provide them this on a printed sheet too. Don’t interrupt or try to help them complete tasks, regardless of how frustrating watching them might be! It’s important that you remain calm, friendly and impartial throughout.

Analysing the results

You should analyse the results of your testing as quickly as possible, preferably the next day.

There’s no need to write a long report, or to cut together expensive videos (if you recorded the tests). Simply get all those involved in the project together and debrief them on your findings. If you need to write it up keep it short and sweet, a bulleted list should suffice. You should concentrate on the issues you found, recommendations on how to fix them, how easy they are to fix and how important the fixes are to your overall strategy. This will help you prioritise the right issues.

Starting all over again!

Once you’ve solved the immediate issues it’s time to start all over again! Studies show that running 3 tests with 5 users will yield better results than running 1 test with 15 users because you are able to test the work you’ve done to resolve the first set of issues.

If you are in the process of developing your website include user testing as part of the development cycle, this will save you making expensive mistakes.

If you already have a website, include user testing as part of your ongoing website maintenance. Testing regularly will keep you in touch with your user’s needs and keep your site relevant. It will also help you to decide on which new features you should implement.

Test early and test often, you can do it.

Further reading to get you started;

• The usability ‘guru’ Jakob Nielson’s fortnightly email:
• Usability.gov - US government website with lots of advice, tips and resources for developing usable websites