Microsoft Office 2004: A Product Vision Study


Usability is a hobby horse that Microsoft likes to ride with every release, and they have made some excellent strides forward. They've invented a lot of useful concepts such as Wizards that have really made the difference. They've borrowed and refined others such as the task bar, which was similar to the dock on the NeXT computer. Unfortunately, there are a lot of other areas where they just aren't very good.

Having really excellent user interface requires singlemindedness of vision, and Microsoft is a very distributed entity. Anyone who believes consistency is the hobgoblin of little minds is not a world-class UI designer. There are many little areas of inconsistency between the applications that would require someone to take totally fascist control in order to eradicate. For example, I can Paste Special from the right mouse button in Excel, but not Word. I can paste just formatting in Excel using Paste Special, but this is done a different way in Word. The list of these little gremlins is practically infinite, and they are annoying. More than annoying, it is impossible to gain an impression of quality from anything that has a lot of little nits. The Japanese car's success was built on this realization, and before Microsoft can really fix their reliability problems completely, they have to address this quality perception.

A second area where Microsoft falls down is that it is willing to mess up its user interface for marketing purposes. When a new release comes out of a product, marketing inevitably wants the new features to be as out in the open as possible. But many new features deserve to be buried deeper if exposing them at the top level pushes aside some standard feature that's used more often or obscures by creating too many choices. MS Word's "Save As Web Page" is a prime example. This should be an option on the standard Save dialog box, not a menu choice of its own. Also, if Microsoft finds itself in a bind, it will choose an inferior UI that can be implemented more quickly and try to market its way out of a corner. Excel's famous "marching ants" cut buffer is a prime example. It doesn't really work the way Cut/Copy/Paste is supposed to, it could have (Quattro Pro's did), but it was hard to do it right and the ants are close enough they can market their way around it. Microsoft are very susceptible to the argument that even though a thing may be better, if it doesn't sell more product immediately, it isn't worth doing. And it's hard to argue that making something right that is almost right already will sell any more product. They consider the 80/20 rule to be gospel: get the 20% right and 80% of users are happy. It's a pity for the 20% of us who are more discerning.

The third and biggest area where Microsoft falls down on usability is they do everything by committee and focus group (aka usability study). This is a time honored tradition that began with consumer packaged goods companies. Steve Ballmer comes out of the fats and oils soap business originally, so it isn't too surprising. Microsoft in fact pioneered the idea of using product managers for software. This is not a terrible approach, but it leads only to incremental and reactionary results, almost never to the kinds of quantum productivity leaps that only true visionaries can architect into a design. A committee can protect you against screwing up too badly, but it can't make you succeed spectacularly. It can, however, lead to spectacular failures. Coke has New Coke. Microsoft has a long list too where usability is concerned. There's the ill-fated Bob. There's the Office Assistant Clippy. The list goes on of ideas that Microsoft thought was great that wasn't because some committee had a big idea and nobody with real user interface design sense could get enough say to shoot it down. Don't you hate menus that claim to learn how to organize themselves by watching you work? Instead, they just seem to reorganize themselves on a whim--choices you could make before are suddenly banished to deeper menus. It's jarring. Any user interface that changes in a non-predictable way is a bad idea.

What should Microsoft do? Create a rigorous and sensitive usability review group that has total authority to enforce consistency across all products. Staff that group with visionary user interface architects. Quit doing 20% of the work and calling it good enough. Quit letting marketing design user interface. Quit letting product managers design user interface. Require the committees to focus their attention on market requirements (i.e. communicating the customer problem or desire, not a specification for a feature). Be prepared to tackle some big areas to make productivity better. Get the engineers to earn their keep. They're all multi-millionaires and still raking it in, yet we're getting limp releases on a lengthy 2 year cycle. Kick ass and take names or get some new blood in that can.

Here is my pet list for Word and Excel of what could be done for a 2004 release on usability. I'm not even going to list all the small inconsistencies in the category I mentioned above, the new UI design visionaries should fix those too. These are my pet problems with the apps. Granted, many of these are nits, but I like the quote "God is in the details" where user interface design is concerned. The marketing committee will argue, "Nobody will care, we won't sell any more units." Guess what guys? You already have a problem driving upgrade revenue. It's only going to get worse. Give people solid value and quality. Microsoft's existing approach is what got American car companies in trouble and gave the Japanese their opportunity years ago. Bigger tailfins are not going to sell these suite upgrades.

All Office Apps

  •   I can search the web on my standard file dialogs with a button click but I can't search my PC. Now I ask, if I'm looking for a document to load into Word, which place am I more likely to want to search? This button is just a shameless plug for MSN put there by marketing. Give me a button that rapidly searches my hard disk, and that will then search some coordinated list of servers for use in corporate environments.
  •  Once very app has notebook tabs (see below), allow the tabs to be mixed. In other words, let me embed a spreadsheet tab in my Word document and vice versa. The current binder is a hack.


  •   Allow notebook tabs like Excel has for document sections. The document map is nice, but this is more efficient of scarce screen real estate. The concept of these tabs should be tied to the current document section idea, without much concern for backward compatibility. To create a new "chapter" that starts on a new page properly users today potentially have to work with sections, heading styles, and page breaks. It should be one simple operation. While we're at it, tie in document templates to this concept. For a lot of large documents, you'd want the template to provide a table of contents tab, a "chapter 1" tab, and an index tab to begin with. In some cases a cover sheet tab that vertically centers the document information would be handy. If these features all work together seamlessly, we get a lot of simplicity, power, and labor savings in one fell swoop, and users have to learn fewer concepts.
  •   The Style Organizer is an abomination. Hard to understand, hard to use, and did you even know it exists? There has to be a better way. On top of this, I have multiple concepts that do similar things. Why aren't themes part of the style system instead of a whole new thing? Templates are yet again another concept unlike themes. These should all be totally integrated. Make it simple. Make it consistent. Make it work.
  •   Make bullets, block indents, and auto-numbering work right. They are mid-way between a good demo and working right. They are useful, but it is very easy to confuse them. Office XP is supposedly better in this respect.
  •   Why do I need "Normal" view? Get rid of it and default to Print layout. WYSIWYG is not a new idea.
  •   Provide a real Outline processor. The after-thought in Word is seldom used and mostly lies buried. Lotus Manuscript was the last word processor to really do Outlining the way it should be done for people who like to Outline. BTW, a table of contents tab, if properly implemented, could be an excellent hook on which to hang outline processing so its fast and easy.


  •   Eliminate "marching ants" and get a real clipboard going.
  •   Do relative copies that work the way Lotus and Quattro did. This is obscure when described but every spreadsheet user encounters it sooner or later if they ever copy between spreadsheets or layers of a notebook. Microsoft simply got this wrong.
  •   Do relative operations in macro recording. 99% of all keyboard macros want to be relative to automate repetitive tasks.
  •   Figure out how to do styles that work as well as they do in Word so that formatting doesn't have to be so manual and piecemeal.
  •   Dump the current charting and start over. It's a total usability disaster that confuses the heck out of even hardened users. There are many simpler ways to accomplish creating a bar chart and making it look the way you want it to look. I would bet that in this day and age of web design and Flash animated graphics we could even make the charts look a whole lot more exciting. If this facelift were done right it would sell the release all by itself.
  •   The Data Filter feature was nice. Now how about letting me sort and do all the other things I want to on each column?
  •   Guess a good file name from the top leftmost text just like Word does for new files.

What about the Internet?

All material 2001-2006, Robert W. Warfield.