A recent trip to the copy room got me thinking about Donald Norman’s usability classic The Design of Everyday Things (DOET). As one of the larger departments on campus, we’ve got an industrial-sized copier that sees pretty heavy use. Like most copiers at this level, they have all kinds of features and functions. One of these many functions is email (e.g., emailing copied documents to yourself or another person).
A couple of days ago, I happened to be in the copy room when a colleague asked if I could help him with this email feature. I was glad to help, but at the same time, mentioned to him that I rarely use this copier and have never used the email feature. Still, the interface was reasonably intuitive and as far as I could tell, it looked like he had correctly input all the necessary information. The only interface detail that seemed to be causing confusion was that the system hadn’t given him any visible confirmation of these correct inputs. More specifically, it hadn’t given him a confirmation message and so he didn’t know if the email had been sent. There was no signifier. So together we carefully searched the touch-screen interface for any hint that the email had been successfully sent, but to no avail. Pressed for time, he shrugged and moved on to more important things. As Norman illustrates in DOET, this type of user frustration is all too common.
So why didn’t the designers include a confirmation message? Did they look to Cooper and design for intermediates? Did they see this user scenario as equivalent to email? Obviously for heavily used interfaces, such as email, confirmation messages are redundant and unnecessary (and it’s got a “Sent” box if the user feels the slightest bit uncertain). But what about situations such as this, where a tool is used only sporadically at best? Does the novice muddle through or hope to find someone nearby who can tell him? Is there or should there be a quasi-decisive metric for including a confirmation message? 50% usage? 60%? Or should it be something simpler (e.g., if it has a Sent box, then no, but if it doesn’t, then yes)? Perhaps some help with this confirmation conundrum can be found with academic research databases in which there are definitely some that still include confirmation messages (i.e., for those instances in which a researcher wants to email article to him or herself). How do the designers of these research databases decide whether or not to include a confirmation message? Do they lean more towards Cooper or Norman?