Blog: Sites and Snowpaths, Part I

Maybe you’ve heard the urban legend about how your college designed the asphalt pathways across the quad.  

At some point, probably after World War II, the college decided to save wear and tear on the quad’s grass by installing asphalt footpaths for the students.  When the administration couldn’t decide how to route the paths, an enterprising engineer came up with a novel solution: wait until the first snowfall.  After the snowfall, patterns of footprints provided a key to the paths that students used the most.  I heard this legend about my alma mater and so did my father about his.  My fellow alum Steve Jones didn’t hear it about our alma mater, but instead about a Cambridge, Massachusetts school that he called Hah-vuhd.

Or maybe the engineer suggested following footprints in the mud after a heavy downpour.  One source I found online said that at San Jose State, engineers looked for paths in the grass.  Fact is, I’ve heard this story a few different ways, but the point remains the same: using data (Aha!  You knew I’d get to data eventually.) provided by footprints, the engineer laid out the best path for the students.  Landscape planners call this concept a desire path.  

Donald Norman, the Norman of the Nielsen Norman usability firm, espoused desire paths for  developing the basic information architecture (IA) for a website (thanks for the tip, Marek Lis).  By understanding where users want to go on your website, you can develop the best paths for them to get there using data and that lofty oxymoron, common sense.

First, the common sense bit.

Years ago, I designed the home page and top-level IA for a website pitch I worked on years ago.  The client wanted a website for a prescription medication that they intended to advertise heavily on TV and in print.  Back then, Google was the funny search engine with the simple page and had not elevated itself above Yahoo!, Alta Vista or Lycos.

Playing with markers and whiteboards, I kept running into dead ends.  It seemed to me that people coming to the site would want to answer all sorts of questions.  I couldn’t figure out which questions had priority over others, the first step in developing an intuitive IA.

Then I had an epiphany.  Or perhaps a beer.  Actually, what really happened is that I got frustrated.  “Who even wants to come to this crummy site anyway?” I asked an empty room.  Then, it hit me.  I knew exactly who would come to this crummy site.

Since the client expected TV and print to drive traffic, all we had to do was consider who would respond to these mass media tactics.  First off, it meant that we had to assume that only people who suffered from the condition treated by the medication would care.  Further, it meant I could rule out two groups of people right away:

1. The ultra-skeptical.  Some people affected by the condition treated by the drug would see the ad, say “yeah, right” and go on back to whatever they were doing.  In other words, the site didn’t need to try to convince the unconvinceable.
2. The convinced.  On the other hand, some condition sufferers might take the advertising to heart to the point where they would call their doctor for an appointment right away.  These people would not come to the site either, so the site had no need to preach to the converted.

What consumer groups did that leave?  Really, just one: consumers who found the advertising interesting but not completely convincing.  This group fell into three basic sub-categories:

1. Those who wanted to know more about the benefits of the drug
2. Those who wanted to about side effects or potential interactions with other drugs
3. Those who wanted to know the science behind the drug

As a result, we recommended a website with four main elements on the home page:

1. Drug benefits
2. Taking the drug (side effects and interactions)
3. Mechanism of action
4. Links for health care professionals

And we won the business to boot.

More importantly, the experience taught me to look for data to help answer questions about what kinds of web content should go where.

In the following post, we’ll discuss what kind of data marketers can use now and where to find it.  But for now, why not take a nice walk?  See what paths you follow.