UXD » Microsoft’s Inductive User Interface

Microsoft’s Inductive User Interface

Sholom Sandalow, Monday, June 25, 2007 @ 3:53 pm

MSDN (the Microsoft Developers Network) has a short introduction to a relatively new trend in the way Microsoft thinks about Interface Design.

Inductive User Interface (IUI for short) is a term that describes the collection of methods and guidelines for designing interfaces that, according to Microsoft, are easier to follow than the current generation of software products are.

According to Microsoft, IUI gained traction as a design process as a result of the research they’ve done on actual users performing tasks on their products. In short, they found that a number of important assumptions that are commonly made by User Experience practitioners are incorrect. They found that, contrary to the commonly held notion, most users are unable to successfully perform even basic computer tasks. The article stated 3 key reasons as to why they have concluded that software is hard to use:

User’s don’t understand the software’s conceptual model. From the original article:

“The interface design for most current software products assumes that users will understand a conceptual model that the designers carefully crafted. Unfortunately, most users don't seem to ever acquire a mental model that is thorough and accurate enough to guide their navigation. These users aren't dumb — they are just very busy and overloaded with information. They do not have the time, energy, or desire to wonder about a conceptual model for their software.”

Even expert users never master common interface tasks. From the original article:

Designers know that new users may have trouble at first, but expect these problems to vanish as users learn common tasks. Usability data indicates this often doesn't happen. In one study, researchers set up automated equipment to videotape users at home. The tapes showed that users focusing on the task at hand do not necessarily notice the procedure they are following and do not learn from the experience. The next time users perform the same operation; they may stumble through it in exactly the same way.

Every piece of functionality on a screen takes effort to figure out how to use. From the article:

Most software products are designed for (the few) users who understand its conceptual model and have mastered common procedures. For the majority of customers, each feature or procedure is a frustrating, unwanted puzzle. Users might assume these puzzles are an unavoidable cost of using computers, but they would certainly be happier without this burden.

Most current software GUI’s aren’t addressing these problems. Instead, assuming the user’s (1) are familiar with standard Interface controls (2) have the time or the desire to learn the software’s conceptual model (3) Are willing to put up with a steep learning curve for additional functionality rather than use a more basic, yet simpler product.

As a result is what Microsoft calls the Deductive User Interface (see image). An Inductive User Interface is one whose screens require the user to figure out what can be done, and how to do it. The more time spent trying to figure out what can be done, the less energy and patience the user has left to actually perform them.

Microsoft’s solution is to design interfaces that induce, or lead, the user through one task at a time. As such, the computer screen should act not unlike an expert standing over the user’s shoulder, directing them through one screen at a time. The four essential ingredients to designing an IUI are:

1. Focus each screen on a single task.
Don’t try to accommodate multiple distinct and possibly unrelated tasks onto one screen. This will potentially overwhelm the typical user, in order to satisfy the expert, or speed user.

2. State the task.
Part of identifying a task, is stating it, clearly. This sounds elementary, but there is actually a lot of literature on the advantages of compact, even terse language in interface design. IUI screen title should use natural language and state the exact task at hand, using verb / object phrases. The example Microsoft gives in the article is from it’s redesign of Microsoft Money: One of the original screen title’s was this too general “Account Details”, whereas the redesigned screen title was “Change account setup”—much clearer.

3. Make the screen's contents suit the task.
Once users have read the screen title they will proceed directly below, to the contents of the screen, IUI’s make that transition effortless, as the tasks associated with the screen are intuitive and natural, corresponding directly with the title (or primary task).

4. Offer links to secondary tasks.
Unlike the Wizard, that ubiquitous and often controversial feature of many a Microsoft product, IUI’s aren’t meant to be modal, and according to Microsoft, aren’t intended to impede the expert user. Adding links to secondary tasks allows the user some flexibility in the way he/she goes about performing their tasks.

Criticisms:
The web is full of interfaces that exhibit many of these same characteristics for a few reasons:

Relatively slow reaction times as commands are frequently sent over the internet and processed remotely

Products need simple interfaces so as to flatten learning curves, to thwart the relatively quick abandonment resulting from the democratic, highly competitive nature of the web

Microsoft calls the IUI design process an extension of the Web –Style Interface, and a few bloggers have commented that there isn’t really much new here. Just a rehashing of tried and true design practices applied to the desktop model.

Additionally, there has been much discussion in the blogshpere on the relative merits and disadvantages of IUI. Most comments have lamented IUI’s similarities with Wizards, which have been rightly blamed for dumbing down the population of computer users by unnecessarily shielding them from any complexity, and preventing them from learning.

user experience design

Comments: 8 so far

Microsoft Bob is resurrected!

The Wizard of Doze wrote:
Focus each screen on a single task.

ROTFLMAO. Let’s all go back to single core processors and application-based context switching (as opposed to pre-emptive) on a 64K Pentium I running Windows 1.0.

I mean, hello Microsoft, people multitask!!!.

The Wizard of Doze wrote:
State the task.

Work is often so complicated that it would take more to state the task than to perform the task.

The Wizard of Doze wrote:
Make the screen’s contents suit the task.

No, give the user the ability to select from lightweight UI’s that can be tailored to facilitate that task.

The Wizard of Doze wrote:
Offer links to secondary tasks.

No, eliminate wasteful mouse clicks and complex paths leading from the primary task to a required sub-sub-sub-sub task option by simply flattening out the possibilities and let me what I want, when I want.

Comment by Marc Clifton, Wednesday, June 27, 2007 @ 1:14 pm
Just 2 cents of wisdom …
IUI guidlines is great thing to follow if you work with users that DO NOT LIKE computers, DO NOT WANT to work with computers… and generaly prefer doing thing the way it’s been done for last few milenia (the HARD WAY).

Comment by Slobodan Ninkov, Thursday, June 28, 2007 @ 5:01 am
Arrgh! Please don’t use apostrophes in plurals!

E.g.:
- “User’s don’t understand the software’s conceptual model”
- “Instead, assuming the user’s (1) are familiar …”

Even with acronyms and initialisms there is no need for an apostrophe: IUIs, GUIs etc.

And “its” as a possessive pronoun also has no apostrophe (”The example Microsoft gives in the article is from it’s redesign of Microsoft Money”)

This sort of thing makes me wince, particularly in an posting that is describing good UI practice..

Comment by Andrew Taylor, Thursday, June 28, 2007 @ 5:07 am
Andrew Taylor is a spelling nazi!

Comment by Slobo, Thursday, June 28, 2007 @ 11:35 am
What’s all this noise about this article it’s dated 9. feb. 2001. Why make all this noise about something that clearly happened about 6 years ago?

Comment by Kjartan, Friday, June 29, 2007 @ 10:35 am
The idea of IUIs may be old, but Vista represents a sharp increase in its influence, and I expect many will be copying the Vista approach without appreciating the underlying theory or its limitations.

There is a place for task-centered IUIs -ATMs and other kiosks come to mind, along with certain use-once web apps. However, the very notion of a task-centered spreadsheet or word processor is terrifying. I can’t see IUIs scaling well to broad, complex, flexible, and dynamic tasks without degenerating into a ponderous and confusing maze of multiply interconnected windows or pages. Also, while task-centered UIs may be more effective in telling users what they can *do*, they’re pretty crummy at telling users what they *have*.

The problems MS cites with object-centered “deductive” UIs really exist, but maybe the problem isn’t inherent in being object-centered but with manner it is implemented. Too often such designs (including MS’s) don’t follow the basic principles consistency, clarity, tolerance, and proportionality. Among other things, failure to adhere to these principles has made users afraid to do anything unless the computer or someone explicitly says it’s all right.

Comment by Michael Zuschlag, Monday, July 2, 2007 @ 8:36 am
All of this is presuming that we can deduce what users want to do. It’s especially ironic that this comes from a company that has shown little real insight about what users want to do in the 21st century. Save us from a future where we are limited to tasks already identified by software developers.

Hopefully this notion is seen as only one lens with which to view a problem. There are always times when we do know enough about what a user is trying to accomplish to keep it simple, and certainly we can work to make smarter and smarter interfaces that can make good guesses about what someone might most want to do next, but Microsoft seems to be taking too specific action in response to their initial observations.

Maybe I just don’t get it, but I remain skeptical of the emperor’s new interface.

Comment by Kevin Steele, Friday, July 6, 2007 @ 10:37 am
“As a result is what Microsoft calls the Deductive User Interface (see image). An Inductive User Interface is one whose screens require the user to figure out”

“An Inductive” here should be “A Deductive”
Probably confusing for people new to the concept.

Comment by BWhite, Thursday, December 20, 2007 @ 3:17 pm