The lessons learned from Howard beautifully correllate with Kathy Sierraâ€™s Quantum Mechanics of Users
If users are asked for their preference, they will likely tell you something very common and rather vanilla. When given the experience of the choices, say during a taste test, the aggregated results are NOT similar, and in fact group around a handful of common choices.
Design Investigation for Software
My take on this is that during Design Investigation for software – be it website, desktop application, even for print – if at all possible, build the DEMO of the choices involved and get the trial user into the experience and THEN ask the user what is their preference.
This approach in awareness of user preference, user experience, and human variability could take the design into the â€œI Rockâ€ space of the passionate user.