We all know when we don't like a design. And many of us might even have strong opinions about what constitutes "good" design. But when it comes to the design of your website, all that really matters is whether or not your users respond positively.
When a design is dictated by the consensus of a committee, or even by the opinions of a designer, the website may still prove frustrating to its intended audience. However, by inviting users into the process, we can remove assumption and guesswork and deliver a design they will love.
Questions that Quantify
A question like " Do you like the look and feel of the website?" asks only for an individual's personal preferences. Instead, well-planned user testing will ask questions such as, "What does the colour scheme tell you about this organization?", which looks to align the design with your mission.
User testing also helps to guard against making the website a mirror of your internal structure. Everything from the navigation structure to the buttons you invite people to click contributes to the impression you make on your users. User testing will indicate how successful the website is in this regard and where it needs improvement.
The methods used in usability testing can differ from project to project, depending on a project's goals. However, here are some typical testing methods that can be applied to most projects:
Aesthetic Survey Test
Decide on two or three words that are important to your brand (e.g., "friendly", "caring", "professional").
Users are shown a set of four thumbnail images — 3 competitor websites and your design — and are asked to "click on the image that feels the most friendly". If your website's design wins the most clicks, then the design is a success. And equally as helpful, if your design doesn't receive the most clicks, then we can work at improving the design by analyzing what caused users to select a competitor's website instead.
Sometimes a project might not have clear competitors. In that case, we will set up the questions differently, asking users to rate designs on a 1-5 scale for each of the chosen brand words. (e.g., "How strongly do you agree with this statement: This design is friendly?")
This is a helpful way to test how well your design leads users to act. The users see the design and are asked a question relevant to your project. "Where would you click to start your course registration?"
The resulting "heat map" will show where users clicked, and will help to indicate where your calls to action need more work.
This simple test focuses on two main points: "How well do users understand the hierarchy of the design?" and "Is the purpose of the website clear?"
A design is displayed for 8 seconds and then taken away. Then the testers are asked:
- What do you remember? List back all the elements you found on the page.
- What do these people do?
If your testers are able to remember the page's key information, then you know the design has prioritized the right things.
Testing Developed Prototypes
The very best results come from connecting directly with users. We can use live testing sessions in which users log into an app which records their interactions with your website while also allowing us to learn from their personal reaction to the design.
Grimaces and furrowed brows don't make it into analytics reports, but they are no less important when designing user experiences.
We can use a series of light-weight tests that are simple to set up, and which can return results in a matter of hours. These methods test:
- How well a design communicates brand values.
- How much of a website's main messaging can be remembered after just a few seconds of exposure.
- How effective a design is compared to your competitors.
Testing your information architecture
If you are at the beginning of a new project, or going through a content audit, card sorting can be a helpful way to make sure that you are organizing content in a way that makes sense to your users' mental models.
Quite often, the more "expert" we become in our fields, the more our "professional" mental models will differ from the typical customer. Card sorting helps us to make sure we are presenting our content in a way that makes sense to the people who will receive what we offer.
Open Card Sorting
In open card sorting, you begin without any pre-determined navigation structure in mind.
Instead, users are presented with a stack "cards" which represent the types of content your website will offer, and they get to decide how to best organize them.
By analyzing the decisions of a group of users (usually around 15-20 will suffice), you will start to see trends or commonalities emerge.
Closed Card Sorting
In closed card sorting, you have already determined the top-level categories or sections of your website's content.
Users are tasked with assigning a more comprehensive stack of cards to your pre-determined top-level sections, or to discard them if they decide cards don't have a natural home.
This process will help you to organize your content in a way that more exactly matches the mental models of the people you are seeking to reach.