All products are made for users. A developer may earn from building the product, a vendor may gain profit from selling it, but if the end user does not benefit from its value – the product wasn’t worth creating in the first place.

This obvious, yet very well-formulated statement comes to mind when you are trying to implement user experience testing (UX) in a company that does not have it.

With this article, we will try to explain simple steps for top managers, to achieve the following:

  • define if their company is ready for UX testing;
  • pick the right kind of testers;
  • conduct the effective and efficient UX research;
  • extract maximum from the research outcome, and integrate UX into their web agency process


Most web production companies wonder why they should care so much about user experience. The clients are buying their services anyway, and the users still use their products.

But this is the ‘blunt executioner’s’ approach. The client says, “Do it like this!”, and the company follows instructions blindly, regardless of what effect this would bring, who would actually use the product, and how it would help their client’s business.

In the end, the product is made strictly by Technical description, but it does not benefit neither the client, nor the developer.

Companies who wish to grow from ‘blind contractor’ to the level of ‘strategist’, from whom the clients would seek both expertise and execution, must dive into the specifics of every project, and actually advise their clients on what is best for them.

To achieve this, a web agency has to make a qualitative leap, and figure out how users behave on the Internet. Or, at least, how they behave when they use the company’s products.


Usability testing (UT) is a study, aiming to define whether a given interface is convenient for its intended purpose. UT is only considered such if performed by real representatives of the target audience.

The respondent is required to perform a certain task within the tested interface (e.g. publish a post or a photo, sign up for a service, make an online purchase etc.), while the host watches the tester closely and traces their behavior patterns (difficulties, failures, assumptions, comments etc.). For adequate results, the interface has to be tested by several respondents.

There are two usability testing methods – qualitative and quantitative. The qualitative testing reflects user impression on the subject website. The quantitative gathers statistics on user behavior within a certain pool of users. For the qualitative method, 5-7 respondents are usually enough. The quantitative method requires up to 20 respondents, but it is only needed for large-scale websites, where the engine has to cope with multiple queries simultaneously. For most interfaces, the qualitative analysis should suffice.

Number of test users vs. usability issues detected ratio


Probably not. Although it is a totally understandable intention. There is a whole market of usability testing companies, which humbly suggest saving your time and money, required for domestic research.

However, outsourcing UX will not help your agency develop in any way. At best, it will only provide you with data relevant to your current project. But in the long run, this gives you absolutely no clue on how to record and predict the user's behavior.

On the contrary, when you set up a UX process flow inside your company, all previous research would contribute to future projects. Only by conducting regular UX research, you will be able to grow, learn and evolve UX-wise.


Start with the basics. Before moving on UT itself, work out project prototypes, User Personas, Use Cases, and make it a rule to test everything on colleagues. If this is made common practice, implementing a full-scale UT will be nearly painless for your business processes.

Many companies wonder, which development stage to introduce UT on. The answer is: as early as possible. Flaws in your interface should not be dragged through all levels of production. The earlier the problem is identified, the cheaper the solution.

The ideal UX workflow is shown on the picture below.

UX testing process infographic


Your testers are human beings, and they are fully aware they are being conducted some research on. That knowledge alone may cloud the experiment’s purity. Hence, your company must create such amicable conditions that would make your respondents feel comfortable during the test.

Below are some knacks of how we do it at Vintage:

  • assign a special ‘testing room’. It should be cozy enough for the newcomer to quickly adapt to and perform the tests in a relaxed state of mind;
  • do not overburden your testers with strict framework and complex terminology. Briefings must be as short as possible, and held in an informal manner. Tea and cookies also help a great deal!
  • one session – one test. Avoid giving too many tasks at once. The respondent’s attention must be focused on the immediate subject;
  • and most importantly, do not reprimand the tester for doing something “wrong”! A user is never wrong. If they fail to achieve something with your interface, it’s the flaw of the interface, and not the user.

As far as technology goes, the best known tool for tracking the tester’s activity is the screen capture. The observer may take notes on the tester’s overall performance, and the respondent, in turn, may comment their actions out loud, but only screen capture would actually tell you what the tester did, when they did it, and why such action was taken.

Eye-tracking is a more complex addition to the screen capture. It lets you see the emotional counterpart of the tester’s performance. This is especially useful in design evaluation. Although, the more technical stuff you have in the room, the harder it will be for testers to behave naturally and feel comfortable.

UX testing, something went wrong

As we can see, technical requirements to most usability tests are more than modest. Basically, a computer with the subject interface, a web camera, and some screen-recording software is enough. The rest lays on the shoulders of your UX staff.


  • Preparation. Define your objectives, and formulate hypotheses for verification. Create tasks, based on the hypotheses.
  • Select your respondents, and set up the testing infrastructure.
  • Conduct the research. Ask respondents to test your interface by completing the formulated tasks, while commenting their actions out loud. Film the process and record the screen capture simultaneously.
  • Analyze your results. Sort and categorize the collected data, and define usability flaws that require immediate attention.
  • Sell your expertise to the client. Draw up a usability testing report, and present it to the client in a form of a slide show, demonstrating problems and suggesting solutions. Remember - no one wants to spend money on fixing the issues. It is important that your report clearly shows why the issues need to be fixed.


Implementing the UX process in your company is, without a doubt, a profitable investment. Sometimes, however, a less resource-intensive alternative may be needed. Thankfully, there are online tools, which can somewhat substitute for in-house testing.

The most widely known of these tools are Loop11 and Usabilla. They let you upload a fragment of your interface, and create a task for a number of anonymous testers from all around the world. After the task is completed, you get a report based on hypotheses you have put in your assignment.

Both services seem to function fairly well. The main difference between them is that Usabilla only works with static images (screenshots), and thus has limited flexibility for interaction research. Loop11 can deal with live HTML, but is generally more expensive. You can try both tools for free, and see their pros and cons for yourself.


If you wish to remain a ‘yes, Sir’-type company, and stay strictly within frames of Technical descriptions, you can probably live without user experience testing. But if you are set out to improve and evolve as a full-scale web production, you will find out that usability testing is not only a tool to check for possible flaws, but also a method to save your time and costs in the long run. A method, which would bring your company to a whole new level of strategical development.