User Research methods: User interviews and questionnaires to understand users

Article Description: How do I Know "she" ?

Last time we discussed the role of UE in the team, there is a point that we have reached a consensus that the user is the basis of all experience, if the user's requirements have not been met, a good experience is naturally impossible to say. So how can we understand the needs of users?

We all know usability testing, questionnaires and other ways to communicate with users, these methods have their own advantages and disadvantages, if the analysis, I am afraid at least three books to write. Now let's put them aside and look at the question from another perspective: what is the way users ' needs are expressed?

For example, a small child is hungry, and he may say, "I want something to eat." And then you know you should get him something to eat, and if he doesn't say anything, grab something and eat it, it's obvious--he's hungry, what if he says "I want to eat hot pot" and you don't have a hot pot but steamed bread? We will explain the problem later.

But you can at least see that the user's needs are communicated in three different ways--goals, actions, statements.

In this case, the user's most fundamental need is starvation (we don't usually need to know the user's most fundamental needs), the goal is to find something to eat, behavior shows this goal, he thinks hotpot can solve this problem. All we have to do is to provide him with the right food according to the information. Here we provide is steamed bread, the children see steamed bread, there are two possible, one is nothing to say, grabbed over the crazy gnawing, the other is a mad gnawing one side angry. The first situation shows that the choice you offer him is more practical than he thought. At the same time, the user said that the truth is not necessarily their real needs, behavior is the most real. The second situation shows that you do not understand the needs of users, need to collect more data, such as his love to eat rice or pasta, like sweet or spicy.

Of course, most studies are much more complicated than this one, but in general, we have to know what the user is doing and why. So you have to combine the various methods and then describe a complete user image.

The user requirements are composed as shown in the following figure. Why does "behavior" account for half of the weight? I personally think that by the Chinese culture of the implicit and the mean philosophy, domestic users may rarely be sincere and accurate to express their ideas, so should be in behavioral research bias.

Let's not think about how to analyze the data, but now just think: what are some of the ways we can collect this data? Look at the picture below:

As you can see, web traffic and log files, as well as the widely admired eye-movement experiments, are used to understand what users do (behavior), while user interviews and questionnaires are used to understand why users do so (goals and assertions), and scenario surveys, usability tests, and CRM statistics are between goals and behavior.

First, talk about user interviews and questionnaires.

The two look very similar, both of which raise a bunch of questions for the user to answer. But there is a key difference between them: quantity. The user interview is a sample survey with a small number (no more than 10 users per type), while the questionnaire is a large scale census. The difference of quantity determines the nature of the two methods, one is the qualitative research method, the other is the quantitative research way. But they are useful for discovering the user's point of view, and you tend to find new ideas that you haven't considered before in your users ' responses, which may change your product's thinking.

There are also differences in the form of operation between the two. User interview form is a more casual way of talking, and should be careful not to mention "right or wrong" (that is, "yes" or "no" problem), so that users free expression. You can have an outline beforehand, but you must not be scripted. In the time also to maintain a certain degree of elasticity, you usually tell users need 1 hours, but if you encounter a good talk about the user, talk about 1.5 hours is also possible, you have to do is try not to let him run too far:). The questionnaire is more rigorous, whether online or offline surveys, most of which should be a quantity-level multiple-choice question, we usually see "you agree with this statement, 5 points very agree, 0 points completely different", is this problem, users can click and draw Shilai answer. Questionnaire also to avoid "right and wrong", at the same time, in order to ensure that users will not be too long to give up, it is best to test their own answer time, generally can not exceed 15 minutes (I answered more than 20 minutes of questions, but that is a few psychological tests).

Here I just want to emphasize that either way, the skill of questioning and the order of the questions are quite important. If you tell the user at the outset that you are ready to develop several new features, and then ask what the user thinks of the existing product, this is a typical self-denial that will inevitably affect the user's perception of the latter issue. I think this is the part that requires the psychological expert to play a role. It is always a matter of wits to excavate the thoughts of the human heart. In a way this communication process is more like the way you get along with the girl around you. You've been trying to figure out why she's upset, but can't ask directly, because you know, she's never going to answer directly. The only thing you can do is to sigh, "How do I know her?" ”。 Maybe she's just angry that you didn't wear the shirt she gave you, but she just said, "You're really tasteless today." "The behavior is to not go with you to any public places, the goal is ... you analyze it yourself."

Web site traffic statistics, log files are used to understand what users do, but usually do not explain why they do so, and similar to the CRM data. So these three best can be combined with the questionnaire to use. By putting a user's click Stream (Clickstream) together with the questionnaire he completed, you can understand the reason behind the behavior. Of course, the premise is that you can capture a specific user's log record and find the same person's response in the questionnaire. Most of the Web pages are seeded with statistical procedures, just as we do today, as we gather as much data as we can. And in the statistics behind the data mining, is a hard and long-term work.

Usability tests are essentially the same as eye-movement experiments, and their limitations are obvious and can only be used to identify defects and barriers to existing products, which can also be obtained in other ways. So the domestic speculation in the usability test, I personally think that the Internet products do not seem to have much impact. This section is skipped.

Situational surveys are interesting, combining user interviews with usability testing. Simply put, you run to the user to see how they operate in a familiar environment, so that you get more data than you do in the lab, and it's important for some environment-related products to do field trips. You can sneak up on a situational investigation or make a deal with the user in advance. But generally speaking, you can see more things without the user's knowledge, although it may sound a little less than a gentleman. At the beginning of the survey, while you're looking at the user's behavior and writing down the questions, this is an improved usability test. When the user completes his daily work, you can show up and invite the user to make a brief visit, put forward your question one by one, and this is a simplified version of the user interview. The risk of this approach is that the user may not want to, or have no time to accept your interview.

The above is commonly used in some user research methods, some examples refer to the most serious book I've ever read, the user is Always right (just as your girlfriend always does), and "The user is always right" is what UCD (USER-CENTRIC design) advocates and insists on.

Author: Angela