VKI Studios is now Cardinal Path! www.CardinalPath.com
Learn more about Cardinal Path

Online Persuasion: How to Develop Effective Messaging

In previous posts, I outlined a number of techniques you can use to make your website more persuasive. However, knowing the techniques is one thing; developing effective messaging is another.

For example, consider Matching Existing Knowledge. It's a powerful technique. But it can only work if you fully understand your target audience and know what their "existing knowledge" is!

Traditional User Testing is not likely to be much help here. These tests are typically task-based. Even if you have an opportunity to ask questions of participants, you won't have much time to probe deeply.

What about surveys or focus groups? Forget about it. Participants — even if they're aware of what their underlying motivations are — will rarely be able to express them clearly in such a setting.

And the personas (customer stand-ins) you've developed? Won't they have the answers? Probably not, as the personas developed in typical user-centered design processes are not nearly detailed enough when it comes to issues of emotion, deep motivation and underlying fears.

Yes, the above techniques might give you a few insights. But they will never give you a deep understanding of your customers.

The only way to get that is through in-depth, one-on-one interviews. Below are a few guidelines to conducting effective interviews:

Recruit Subjects From Your Actual Target Group

In traditional user testing, we sometimes say "any test is better than no test". And if we don't have a subject available who exactly matches our target user, we'll test on someone who's "close enough". Usability problems, where they exist, will probably be revealed anyway.

Unfortunately, this "close enough" approach does not work when it comes to persuasion. You're looking for the nitty gritty of what makes your customers tick, and only your actual target group can tell you that.

Interview Lots of Subjects

Again, the process needed for persuasive web design is more demanding. In traditional User Testing, you're looking for errors, sources of confusion, etc. A relative handful of subjects (say 5 - 10) will identify all the major ones.

With Persuasive Web Design, you're looking for patterns. You're looking for insights you can generalize to a wider population of users. And these insights are harder to find. So you need to interview more subjects. Probably around 20.

Make the Subjects Feel Comfortable

You need to uncover deeply personal information, things the subject probably isn't even aware of. This is not going to happen unless the subject is comfortable and trusts you. A few helpful things to do are:

  • Dress casually
  • Hold the interview in a quiet, comfortable room with subdued lighting
  • Establish rapport with a few easy-to-answer questions
  • Don't make too much eye contact (it's hard for subjects to be introspective when someone's looking them in the eye)
  • Look and act relaxed (slouching is good!)
  • Don't be judgmental

Focus on Scenarios, Zero in on Feelings and Beliefs

The interviews should focus around scenarios (for example, "applying for a loan online"). But you don't actually ask the subject to perform the task. Rather, you ask how he'd go about the task and — most importantly — what he believes and how he feels at each stage of the process.

Probe Deeper!

Don't be in rush to get through the entire scenario. Probe for more detail, because that's were the useful nuggets of information will lie. You should find yourself asking things like:

  • "Okay... Tell me more about that."
  • "Mmm hmm.... What else?"
  • "So you're saying...?"
  • "Oh I see. And why is that?"
  • "And how does that make you feel?"


A properly-conducted interview can uncover a mountain of useful information. Deep information about how your customers think and what motivates them. Next week, I'll discuss how to analyze this information and put it to use.


Comments (Comment Moderation is enabled. Your comment will not appear until approved.)
.