I’ve designed and sent about a dozen surveys. I’ve helped companies figure out what their clients want by asking the right questions. I’ve gleaned some incredible information from these surveys and it’s helped guide my own business by understanding what people want. But some times what people want, or what people THINK they want, and what people NEED are entirely different things. This post is about what happens when surveys attack.

I recently began working on a website for my church. While I believed that I completely understood what the parishioners needed from their new website, I’m also fully aware of my own ego and I didn’t want to make assumptions about the users. So I set up a survey and started asking some questions. The answers, boggled my mind.

The main reason that we’re redesigning the website for the church is that the current site…lacks…well, it’s sort of…you can’t…ok. It sucks. I don’t blame the designer of the original site. At the time, he was building based on the tech available and it was solid but it’s no longer feasible based on todays standards, for a variety of reasons. The biggest issue I have with the site is navigation. There are 9 menu items each with multiple sub-menu items and 7 additional sub-sub-menu items leading to more than 43 pages. 43 pages. For a church. Some of the pages have less than 50 words of text.

So on a website where navigation and ease of use are poor, one would expect that when you asked, “what do you think about this website (in much better words)”, the most common answers you might expect to receive would be things like, “the navigation sucks” or “it’s hard to find things”. But based on the title of this post, you might imagine that these were not the answers I received. While a few people pointed to these issues, most people seemed oblivious, if not PLEASED, with the navigation and user experience.

This got me thinking. If people like the way the site is designed, maybe we don’t need anything too dramatic when it comes to a redesign. Maybe we need 9 menu items leading to 43 web pages. Maybe it just needs to be responsive so that people can view the site on various devices. Maybe this website is right and I’m wrong. Actually, no.

Henry Ford supposedly once said, “If I asked people what they wanted me to build, they would have said a faster horse.” Ignorance is bliss and I’ve come to realize that a good deal of our parishioners just don’t have any idea what good web looks like. That being said, maybe a lot of people don’t know what good web looks like. For me, good web is about three things.

  • Easy to navigate
  • Great information
  • Hits you in the feels

Now the first two are pretty obvious (I think). But that third one. What do I mean? Well, what I mean is that great design is inspiring. When we look at a dress or a suit or a building or a car or a watch or glasses or shoes that are well designed, we have an emotional reaction. Sometimes great design can mask the fact that there really isn’t any substance but in most cases, great design USUALLY means a good product.

But let’s get back to surveys for a minute. There’s something really important about surveys and that is that you either believe in your survey or you don’t. What tends to happen when people are using surveys to collect information is something called the confirmation bias. Basically, people look at the data that supports their point of view and more or less ignore the rest. “See, right there people said they liked THAT thing.” Sure. But over there they said they didn’t. So you either believe in the data or you don’t. So with the case of this survey that I sent out, I’m left to make a decision. Do I believe in the data or do I not?

I don’t.

I think there were some interesting points that came from the surveys, specifically some of the ones that I wouldn’t have really thought about but for the most part it plays out like a group of people who don’t know much about the internet, talking about the internet. Instead, I’ll follow my three rules; easy to navigate, great information, hits you in the feels.