I was reminded of the above line from Spinal Tap while reading an article in the October 2005 Technology Review called, "The Customer is Sometimes Wrong." It's about Salesforce.com, the company which does web-based Customer Relationship Management software. In case you're not familiar with it, this is software that allows sales people in companies to track contacts, leads, and manage accounts, and it's really a very big industry. Traditional CRM systems are sold by companies such as Oracle, IBM and Siebel, and can be, as the article says, "complex, expensive and temperamental." Marc Benioff left Oracle to start Salesforce.com, and he had a vision to remove all that complexity, and the attendant annual costs of servicing and equipment and upgrades and training, and shift everything over to a web-based application with the data stored remotely.

This meant that instead of all the data residing inside the computers belonging to the owner of the software, it would now be stored by Salesforce.com. It's the difference between using an email application like Outlook, and using Hotmail - one is kept local, the other remote. This caused, to say the least, a lot of skepticism, since much of this data was a) very sensitive if it were to get in the wrong hands, and b) very critical to have access to at all times, for large companies had sales people all over the world that needed it 24/7. Buyers of CRM software wanted 100% control of their information, 100% of the time, even though it meant a lot of pain in other areas.

How do you decide that when a customer says something is stupid, they actually will think it's clever when they see it, and vice versa? 

Salesforce.com made a very conscious decision to ignore what its potential customers were saying they wanted.

They were not the only ones pursuing hosted approaches to data: VC's in Silicon Valley were funding a lot of companies doing similar things. They were telling the companies they were investing in to give customers options, hosted and non-hosted. Benioff didn't buy it, and made a bet-the-farm decision to go purely hosted. As the article says (emphasis mine):

"Benioff and his small staff - working mostly out of Benioff's home - considered it a bad idea to grant their customers the in-house option. So they did the unthinkable: they walked away from the prospective investors who insisted on the 'choice' model."

On the face of it, this goes completely against the grain of user-centered design, where it is considered arrogant to say you know better than customers what they want, and you should get as close to them as possible and understand their every need. As it turned out, Benioff was right, and the customers (and the common wisdom) were wrong, and many of those same customers are now subscribers to Salesforce.com.

How do you decide that when a customer says something is stupid, they actually will think it's clever when they see it, and vice versa? For someone like me who does a lot of user research, this is the sort of thing that causes sleepless nights.

I saw a talk years ago by the late Rolf Faste, who was a professor of design and engineering at Stanford. He drew a bellcurve on a chalkboard and talked about how the quality of ideas follows it. Most are pretty average, while some are very clever and some are very stupid:

Hopefully you only purse the cleverist ones and leave the stupidest ones behind. But how do you tell the difference? The difficulty is, the line between stupid and clever is not stable.

Here's an example: In the early days of the Palm Pilot, they did usability testing to see what features people wanted on it. Palm thought that wireless email would be the killer app. In fact, it was a bomb. No-one wanted it. At that time, people got very little email and they checked it only a few times a day (this was way back in the early 90's), and having it wireless seemed totally, well, stupid. (There's an outstanding interview with Palm's software architect, Rob Haitani, in the book Information Appliances and Beyond, that has many nuggets of insight about creating great, usable, useful products that applies across all product types. Highly recommended.)

Five years later, RIM launched the Blackberry to do exactly that same task, and they had a huge hit. Very clever. (Assuming they don't get shut down by the patent infringement suit, which would give a lot of people a very bad day.)

This brings us back to Rolf Faste and his bell curve. What he did next was to extend the curve into three dimensions, which makes it look rather like a UFO (I've recreated his chalkboard drawing in a more modern medium):

What becomes evident, then, is that in fact the stupidest ideas and the cleverest ideas can be one in the same! What makes the difference can simply be time: ideas that make sense in one context at one time, become really bad ideas in another context at another time.

Faste then drew a number of other UFO shapes around, like they were flying over a landscape, illustrating how these groupings of ideas can all be floating around, moving through time together, increasing and decreasing in stupidity or cleverness, merely by the progression of time:

This doesn't help answer the question of how to tell which side of that fine line between stupid and clever you're on, but it should at least prompt us to keep an open mind. In the case of Salesforce.com, such open-mindedness has taken their share price from $11 at their June 2004 IPO to $40 today, and given them almost a 10x increase in customers over the last 4 years, while Oracle and Siebel's CRM businesses have struggled.