How to Start Customer Research

Sep 14, 2005

"Google has made home pages virtually irrelevant."

So said one of our moderators at our recent LabFest in New York, where we observed dozens of customers using dozens of websites.

The moderator, admittedly being a bit provocative, had observed several respondents entering websites through Google search results, rather than by typing the URL into the browser and entering "properly" through the home page. Often the search results link pointed to a destination page in the middle of the website, which caused the customer to miss the home page altogether.

To be fair, we also observed customers in labs entering the home
page of the website - but even then, many began by searching Google
for the name of the company they wanted to visit.

Stated another way: many users, when not directed how to start, begin their sessions by going to Google and searching for what they want. (A small minority use Yahoo, and almost no one uses any other search engine.) Some companies, depending on their size and popularity, also have a fair percentage of users who do type the Web address directly into the browser.

The key, though, is how to find out: by not directing users how to start the session. In our non-directed listening labs, we ask customers to use the Internet in the way they normally use it at home or work. While we do have a goal for the research, we try to let the customers lead us to the answer, rather than the other way around.

Non-directed research is a foreign concept for some practitioners. On the issue of Google usage, for example, many companies never learn how users naturally begin their site visits, because the usability tester begins each session by directing the user to start at the company home page (or whatever specific page they want to test).

I've always thought this to be a strange way of observing customer behavior - by commanding the customer to act in a certain way! Not just how to enter the site, but what action to take next; what tasks to accomplish; what features to look at; what questions to answer.

Now, directed research is valuable in some tactical circumstances, such as the testing of a specific feature in an online application. But I see that as an exception to the rule: letting customers lead, in a non-directed environment, elicits much more accurate and insightful feedback.

Customer experience is mainly about understanding, and serving, the key unmet needs of the customer. This is a strategic issue that's poorly addressed with a tactical research method. You simply can't find out the customers' priorities if you give them a list of pre-written tasks; there are too many assumptions built in. Instead, why not just ask them to show their experience?

If these ideas are intriguing, a good way to try them out is to see how customers enter your site - naturally. In your next user test, don't force customers to start on a specific page - ask them to show how they get to your site.

P.S. For what it's worth, a quote from one of the LabFest attendees, having observed non-directed research for the first time: "[Our company] saw things in one afternoon that we did not see in a year of task-based usability sessions."

• IN: Column Archive
• #