All products should be tested by the people that will use them, regardless of whether that be your clients’ customers or the general public. You can be the best designer in the world, or think you know your users inside out, but you still won’t know how your users will behave until you observe them using your product.
User testing is the process of evaluating how your product performs for your users. This enables you to spot problems throughout the development process. There are several types of user testing you can do at various stages of development, and this article will discuss and analyze three in more detail:
- guerrilla testing
- remote testing
- expert reviews.
User testing doesn’t have to be expensive. While you can spend a lot of money on laboratory facilities and third party expertize, to test your product and get a lot of valuable insight, usability problems can also be uncovered in a much less formal environment, using cost-effective and simple techniques.
The selection of users is important to get right, as it gives you a more valuable output. They should be closely matched to your real target audience – use any data that you have on your users to determine this. Also consider those who you wish to aim the product towards in future. Think beyond simple metrics such as age and gender, and include factors such as technical competence, familiarity with certain software, or hobbies. Sometimes you may want a more specific target audience of testers depending on what you want to test, while other times you will have a much broader range of testers.
Guerrilla testing
Guerrilla testing is the cheapest and easiest style of testing to set up. The word guerrilla in this context describes the ‘in the wild’ style of testing, meaning that it can be held anywhere. Common places for this type of testing are coffee shops, offices and university campuses.
Participants are usually the general public, colleagues, or friends and family. You normally only need a laptop, or device, and a prepared list of things you want to test. It usually takes 5-15 mins to quickly gauge opinion on a particular aspect of your product. You may also choose to record the sessions (with permission from the person you’re testing).
This form of testing is quick and inexpensive. The result is a list of common usability problems that you can immediately iterate the design with, and that suits an agile development process. It’s advisable for someone other than the designer to carry out this testing, to reduce any confirmation bias and bring a fresh perspective.
Remote/crowdsourcing testing
In order to get a broader spectrum of participants, you might consider using people from a wide geographical area, and carrying out testing remotely via software such as Skype. The easiest way to carry this out is through crowdsourcing methods. Crowdsourcing, in this context, uses a large community of users to test your product. There are many services that provide such a solution and have a wide network of testers, for example usertesting.com and UsabilityHub. You can simply upload a design or prototype and get fast feedback, usually in the form of a heatmap, or the recording of a test session.
A global community means you have a much more diverse set of testers to choose from. You can usually narrow down your audience by specifying simple metrics such as age, location, and language. Some services will also allow you to define more specific requirements.
As well as having a broad reach, remote testing through a crowdsource provider is relatively cheap compared with outsourcing the testing. It’s also easy and quick to set up, and results usually start to appear within a few hours. You will need to leave some time to analyze the results and identify the common usability problems, so that you can iterate your design.
Case study: Account improvements
In Interact, our digital banking platform, we recently made some layout changes to the way we present information for motor finance accounts. We utilized crowdsourcing to test our designs regularly and iterated them until we had something we were comfortable with and were confident would work well for our customers.
One of the methods we utilized was to use a series of dynamic cards that pointed to various ancillary information and tools around motor finance, such as viewing your payment plan or carrying out a check on your PCP mileage. We wanted to test a couple of visual design options, one using text and iconography, and the other using text and photography.
In order to test this, we carried out two ‘click tests’. A click test is a static screen that can be shown to a group of users (in this case crowdsourced), who are then asked to click to choose a certain area. The question is most often framed, ‘Where would you click to …’. In this case, we asked our users to click the place where they would expect to make a transfer. Doing this across several variations with different groups of users allowed us to have a comparison. In this case, there was a significant difference in accuracy, with the iconography performing much better than the photography.
Another series of tests we did was around primary actions on account screens. We wanted to see if users preferred a clear, up-front, primary option, or if a cleaner layout suited them better. A simple preference test, asking users to choose the one they thought more understandable, gave us some guidance.
A preference test shows the same group of users several options to choose from, with a question to guide them – eg “Which is more understandable?”. The order of the options is randomized so as not to bias the results. Added to some additional click testing and guerrilla tests, we had a fairly good indication what worked best for our users.
There are a variety of third parties who provide user testing as a service. The usual process includes:
- An evaluation of your current product, in order to generate test cases.
- Recruitment of test participants using a defined set of criteria.
- Laboratory based testing with a moderator, and a set of observers in another room watching the testing (usually via monitors and/or one-way glass).
- Analysis and reporting of the results by way of a report, presentation, or workshop. Major problems are discussed and solutions advised.
Using external experts to analyze your site gives you the advantage of an unbiased, expert review and solid recommendations. It is, however, a more expensive user testing option, and is best done on a near-finished product in the final stages of development. Observing people using your product provides invaluable insight into what users are thinking when they use it.
Conclusion
User testing is vital if you want to know whether your design enables the best use of your product by your target audience. Without it, you’re designing blind and will only find out when it’s too late. Testing at regular intervals throughout the process will enable you to reduce the risk of usability problems in your final product, and will give you the most customer insight.
Use a combination of remote and guerrilla testing at early stages, in order to iterate design quickly and avoid going down the wrong path. Once you have a near-complete product, you should also consider some expert third party testing to fully analyze your product as a whole and test it with your defined audience.
Want to know more?
ieDigital can help your organization deploy a digital self-service platform that gives you the foundations to foster a relationship with your customer and, in time, extend and flex your value proposition. We can also work with you to better understand your customers’ needs and to prototype, test, and try out new concepts and ideas with them.
For more information, please click here to request a chat or demo.
