Anticipatory design done correctly

(This posting was written in November of 2007 and posted finally today.)

(Beware of some car-speak.) I was looking for a specific part for my car because the coolant was leaking from the thermostat plug to the ground. This part was an o-ring and it fit around a plug. The problem was that I didn't want an OEM (original equipment by manufacturer) part as that part is defective and doesn't take into consideration temperature variances, hence the leaking. So began my search through the forums and I took some keywords to Google where I found a supplier. Here's the website: http://www.mcmaster.com At first glance, I wondered if I was in the correct place. But what I've noticed in many e-commerce parts sites is that functionality is high, given that you can find the part, and the usability is somewhat low. That's evidenced by sites like NAPA, AutoZone and AdvancedAutoParts. 99% of the time, you have to go to the store to get the part for a specific vehicle. If the part isn't online after executing a search (especially on obscure parts), it's either not cataloged or the system is incomplete.

Taking a look at the McMaster website, its purpose is different. Among all the thousands of different mechanical parts, hardware and software, the user begins by generating a scent using the search functionality. However, this is no ordinary search engine. At first, I wasn't sure what to type in, so I just entered the part number I found. Results came back positive and that was easy. But the kicker came in my second search where I entered "battery terminal covers". And immediately, anticipatory design kicked-in. A list of suggestions came up for me to select. There were no part numbers or items without pictures. Upon selecting the correct category, part numbers and their corresponding part numbers came on-screen.

Now THIS is what e-commerce should be! I think those other automotive vendors could learn something from McMaster.

Adaptive User Interface

I've just downloaded and installed the trial version of MS Office 2007 with the "ribbon" interface. In just a few minutes of playing around with it, I've learned that this is probably the first interface I've come across that has an Adaptive User Interface (AUI).

By adaptive, I mean the interface (in both high and low levels) actually changes based on the context of the activity involved. The actual menu items are not user-controlled - instead they are based on user behavior and their tasks. For instance, additional first-level choices appear when the user is performing, say, a table creation or edit, or a picture or art object creation or edit. Specific functions also change based on the object that is selected for either formatting or editing, resizing etc..

The idea is to have the system use anticipatory design practices to better predict user behavior and flow with the daily interactions. Whether or not the system becomes smarter with use is up to the time factor.

Taken this further, I can definitely see an AUI that has algorithms that actually anticipate user actions before they happen. But how could this be without digging into the human brain? This is where science fiction becomes science fact - and it seems we're getting closer to this every day.

As for actual AUI's being in use, I will have to research this subject more to give me my perspective.

** Okay, so by Adaptive, you can say the interface is adaptive if there's personalization functions (displaying or hiding toolbars) as opposed to what I'm thinking in full anticipatory design model, with use of perceived artificial intelligence.

The Impact of Design

After reading Donald Norman's first chapter of his new book, I realized a couple of things:

1. What I'm designing right now has no catastrophic consequences if things go wrong for the user or the system;
2. What I'm designing can be referred to as "low impact" because of the nature of the product (administration). Functionality is more driven than usability at certain times.
3. The only thing I can affect is the bottom dollar, directly dependent on learnability.

Yes, I think in fundamental ways. I can also tell you that I've designed systems or products where catastrophic failures can occur. When this happens, more and more scenarios go through my head to anticipate every situation imaginable. And from Donald's new chapter, it can only come from the designer's head - that is why having usability analysts are important.

Anticipatory design is what we do. Solve it before it becomes a problem. But then that means automation in certain respects - by autofilling users' previous entries, by autocalculating cruise control speed on a car in approaching traffic. The objective is to meet the users' needs even before they start to complain - like a decent waiter at an exclusive restaurant.

While I won't be saving lives (directly) with what I'm doing (as opposed to developing airplane systems or auto-cruise control on cars), I will be improving some people's attitudes by creating a more pleasant and usable interface. Not that administration work is in itself pleasant, but perhaps then, less evil.