Internet Marketing - Matt Bailey [41]
Conduct Tests with Real People
The fastest, least expensive, and most honest opinion I can get is from my mother. Maybe yours is the same way. I find that when I show her a website and ask her to perform a task, she has no problem letting me know about any problems she has. She grew up in a generation without computers and websites, but she has a disposable income, which makes her a target for many companies. However, if she can’t find your site or use it, you have the problem. She doesn’t. She knows how to easily find another website that might be easier.
Watching someone use your website and asking them about the experience can be a moment of clarity that many marketing managers need. Website owners can gain a significant understanding of how searchers view their website, simply by having a third party, disassociated with the website or the development project, use the website.
I find that too many website owners have a “proud papa” complex and believe that their website was created to do the job. However, when it is a site that you see every day and when customers call because they cannot find things on the website (or much less find the website), you have a problem.
Start with a Task
The simplest and most effective test is to set a clear goal for your testers. Simply showing someone a page on your website and asking what they think is not only ineffective, but it’s out of order. That’s not how people browse websites. They go with intention, so measure intention.
Find a specific task, such as purchasing an item or finding a specific bit of information. If there is a process involved, such as filling out a lead form or a purchase path, add that to the task list. If you want to raise the level of difficulty (for you), find additional test subjects and test their path through the site with different starting points. After all, not everyone enters your website through the home page.
Measure the time, observe the mouse movements, and, above all, do not help! You can ask what they are thinking or searching for. You can ask what they see. You can ask what they expect to see or ask whether anything is missing. But restrain your desire to help or point out information. Otherwise, you won’t learn anything.
Simple testing like this is cheap and easy and will result in some very basic improvements. Once you see how that works, then you can move up into the rankings of usability audits, eye-tracking, and high-end user testing. But until then, find the easy stuff and start there.
Set Up a Persona Walk-Through
A very effective method of assessing your content is to see it through someone else’s eyes. By considering another person’s abilities, needs, and background, you can also develop the skills necessary to evaluate your website on an entirely different level. This method begins by establishing a typical user persona. There may be three to four actual persona types, based on your audience, but if this is your first time, just start with a single persona. I described a persona in my example a few pages earlier. The target audience of the website was affluent, retired women who wanted to travel. After the basic demographic information is developed, we now develop the persona.
I find an effective method of doing this is to go to a stock photo website and find a picture that fits the person you are describing, ideally in the context of your target audience. If your target audience consists of engineers who are drawing specifications and they need to know about your product, then find the demographic of your engineer, probably sitting at his workstation; surround that person with context so that everyone on the project knows the persona.
Next, give your persona a name. This makes it personal. You may even find yourself saying that “ Mildred wouldn’t like that!” Everyone knows that Mildred doesn’t exist, but giving your persona a name makes them more human, more understandable, and sympathetic. Programmers have a tendency to consider website visitors as stupid, but once a persona is established, there tends to