Other Research Methods
Although heuristic reviews and observational customer research are the most common methods, there are many other types of qualitative customer research. This section outlines some that are popular among online businesses.
It should be noted that two methods that remain popular today—focus groups and usability testing—do not figure in this section because they are not recommended. Focus groups tend to place customers in the role of product designer by explicitly asking them to give advice on design decisions, which is difficult for people to do when such decisions are not their area of expertise. Additionally, focus group participants tend to try to say the “right thing” to please the researcher or they succumb to groupthink.
Traditional usability testing asks customers to complete a predetermined list of online tasks based on actions the business would like to see completed rather than allowing customers to show the business their natural online behavior. Even though customers may be good at figuring out how to perform a task when asked, doing so takes them out of their typical frame of mind when interacting with websites.
Instead, the following are several tried-and-true methods for understanding customers through qualitative research.
Ethnographic studies are a special form of observational research that take place where the customer would normally interact with the company, such as at their home or office, or at the location of the business itself, so that they are more likely to behave the way they normally would. Just like observational customer research, these sessions are most effective when they are unguided and researchers focus on active watching and listening.
Kenyon Rogers, Director of Digital Experiments for Marriott International, said that his business regularly conducts ethnographic studies at select hotel properties. For example, in 2013 the company piloted a program that allows guests to “use their smartphones to check into the property and open their room door without needing to interact with a Marriott team member.” Rogers added that very soon, “guests will be able to control their entire experience, including ordering room service, extending their stay, ordering transportation, and booking meeting rooms through their smartphones.”
Businesses place surveys on their actual site or app, or email them to customers. Large surveys can have statistically significant sample sizes, but the researcher must be on the lookout for data not representative of the larger customer base due to self-selection bias. For example, not every customer wants to fill out a survey, and those who do may have the strongest positive or negative opinions.
As with all forms of qualitative research, the more open-ended the survey, the better. Surveys that ask customers about specific design decisions place the customer in the awkward position of being asked to provide advice outside of their area of expertise. Rather, understanding whether customers found their overall experience to be positive or negative and providing an open-form field for customers to write about any aspect of the experience they choose can often lead to actionable data.
Customer panels are a subset of surveys: They typically consist of thousands of participants who have elected to give survey feedback on a regular basis. Panels may be run by a company’s research team or by consulting firms on behalf of many businesses. Like surveys, customer panels can provide statistically significant sample sizes, but it’s important to understand the segment of participants being queried. For example, although panels consisting entirely of self-selected users of one business might not be representative of the entire population, they can give the business insight into the behavior and opinions of their more loyal customers.
Eileen Krill, research manager at The Washington Post, oversees customer research for all of the business’s print and digital brands. A panel of about 7,000 customers is included in the many types of qualitative and quantitative customer research she oversees. Krill will ask the panel “a wide range of closed and open-ended questions, depending on the objectives of the survey,” including “satisfaction rating questions.” She points out that “open-ended feedback is generally far more meaningful and actionable than the score itself” because it can help to “reveal the reasons behind the scores.”
One question she often asks the panel is “whether a new product or feature will improve the customer’s impression of The Washington Post brand.” Although in most cases participants say such additions would have no impact, Krill still asks the question in case it provides an important insight. For example, she said, “We once tested the idea of starting an online dating service for Post readers, and that got a lot of people saying they would have a lower opinion of the company.” Krill noted that “I think that research was one of the key things that may have killed the idea.”
Diary studies consist of a business asking customers to take notes and regularly send them back to the company, usually over an extended period of time. These studies may ask participants to take notes only on a specific topic area, like their regular interactions with a new site or app, or they may be more general and simply ask customers how they spent their day.
Google Search Lead Designer Jon Wiley shared an example of an ongoing diary study being run by the company. Through a mobile app, the study regularly asks participants to reply to the question: “What is the last bit of information you needed to know?” The information can be related to any aspect of their life, not only the material they were looking for online. Wiley and his team then “look at the needs that people have in their lives” and try to answer the question, “Is there a way that we, as Google, can find a solution for them?”
A technique used to gain insight into how to organize content, such as ordering and grouping similar navigational links, card-sorting directly involves customers in the design process. Customers are provided with a stack of cards containing information and asked to perform a task, such as organizing them into logical categories. Although it is typically performed with index cards or sticky notes, card-sorting can also be performed online, and there is no limit to the number of participants.
Caution should be observed, because this technique is more guided than the aforementioned qualitative methods and may place customers in the position of being asked to act like a professional designer. However, if performed with minimal prompting, card-sorting can provide designers with a rare opportunity to gain insight into how customers think about information architecture and content hierarchies.
Nate Bolt, design research manager at Facebook, said the company used cart-sorting as one of many qualitative research methods to inform an ongoing redesign of users’ Facebook News Feeds. After recruiting users, the Facebook Design Research team printed out each user’s feed, up to the minute, on paper. Then, he explained, “they would cut out their feed stories, place them on a table, and group them into what they considered to be like-minded categories. That helped us reprioritize the ways that stories are grouped and organized within people’s News Feeds in a real, human way. Obviously, this went hand in hand with all other data, including the system (analytics) data.”
Feedback forms on websites, or email addresses for user feedback, can be a good source of information. Users who provide open-ended feedback in this way often have timely insights into customer frustrations, as well as positive comments on great experiences with the website or the business’s team members.
John Williamson, the Senior Vice President and General Manager of Comcast.com, explained that “One of the best sources for information I have is the ‘Website Feedback’ link” that appears on the bottom of each page on the site. Williamson noted, “I was in banking in the mid-1990s, and I remember how excited we would get if we received a letter from a customer. We really would.” Every day Comcast receives “hundreds of letters” from customers through the feedback link, and because they are a “huge benefit” that helps Williamson and his team understand their customers and their business, he added, “I read every one of them.”
Acting on Qualitative Research
After the heuristic and other qualitative research sessions have been conducted, researchers summarize the insights they have garnered, as well as any remaining questions about customer behavior. For each insight, they write down possible ways of verifying the observations through analytics data, as well as any ideas for testing. This list will be used as a guide when crafting the Optimization Roadmap.