The services of 10 usability firms that I have experienced and surveyed over the past 13 years, as well as my own past practice as a usability expert, and the three standard treatises on the subject by Jacob Nielsen, Donald Norman, and Alan Dix et al., are all heavily influenced by two root metaphors (to borrow this concept from Lakoff and Johnson), whose effect—in terms of metaphorical consistency—helps to explain why the services take the approach that they do, and why it is inadequate for most non-trivial websites or software.

Metaphoric Families that Influence Usability Services

A Website is a Machine. This family of metaphors includes

■ User-Interface is a Control Panel.
■ Buttons, Links, Fields, etc., are Controls (levers, gears, switches, & dials)
■ Interacting with a Website is Using a Machine
■ Usability is Efficiency of Operation.

A Website is a Place. This family of metaphors includes

■ A collection of hyperlinked information is a Site
■ A User is a Visitor
■ Users are Traffic
■ Finding information on the website is Navigation
■ Efficiency is Ease of Getting Around
■ Displaying a web page is Going Somewhere (to the page)

In the general case, the use of metaphors has great influence over the experience of someone who uses them. Perceptually, it is very influential in determining what you notice—the differences the make a difference to you, and therefore what you pay attention to and what you ignore. Cognitively, they have a great influence over how you construe the things that you do notice.

Usability firms and experts employ these two metaphorical families widely in their services without realizing they’re using them. As explicit metaphors, they serve to influence the perspective that these people take when they analyze and assess a website. Metaphoric consistency serves to keep them in these families, so that the use of some of these metaphors leads to more and more use of it within these two families, thereby coloring the way they think about and talk about websites.

Simple vs Complex Purposes

If we go deeper under these root metaphors, we can see what is really going on, and it comes down to this word “Use.” In talking about our purposes and the means we employ to pursue them, it is important to make a distinction between “simple” versus “complex” purposes.

A simple purpose is one where you can directly achieve the purpose by action or using a tool : for example, using the toilet, or using a door, or using a hammer to hang a picture frame.

Complex purposes are those where this 1:1 correspondence is lacking, where there is a 1-to-many relationship between the purpose and the activities that roll up to it. For example, if I ask, “What are you doing?” and you reply, “Well, I’m using Excel,” that would be an appropriate answer if my question was about a simple purpose, such as, “what application are you using?” If I want to know for the sake of what are you are using Excel, then I am in the realm of complex purposes. Are you creating report, calculating payments, analyzing stock prices, or what? Here the purpose is complex. Using the tool (Excel) does not directly accomplish the purpose.

To connect this with the ideas we began with, metaphoric families based on the word “use” are suitable only for contexts where the purposes are simple. This is because “use” suggests “simple effectuation,” not employment for a purpose. The problem is one of a blinding emphasis—making “use” central, as opposed to what end.

When usability practitioners think in terms of “use,” they are assuming that what they are assessing is used only for simple purposes. For a website, this means that what a visitor is doing is “using the website.” Well, that will only be appropriate for analyzing and assessing the website, if what the user is trying to do is simple. If not, it will be very misleading. That’s how you get to the bottom of what’s wrong with the field of usability, and with a lot of UX in general.

Both metaphorical families, Website as Machine and Website as Place, are only consistent with simple purposes, which is why they are inappropriate ways of thinking about most websites. In the case of Website as Machine, the website is construed as something like a power plant, whose ultimate purpose is taken for granted. The only issue the operators have is how easy it is to control. In the case of Website as Place, the website is conceived of as a road system laid out with junctions and signs, where the drivers’ purpose is merely to get to their destination, but what they’ll do when they get there is taken for granted.

Implications for Usability Testing

Usability Testing is typically only adequate for applications whose entire reason for being is a simple purpose or set of simple purposes. This is because the simple task-level is all that is ever tested, and only applications with simple purposes have a high correlation between ease of functionally control (”use”) and their true purpose. For example, a calculator can be evaluated this way, but not accounting software. This is why so much usability work is really just putting lipstick on the pig. A test can result in great scores, and yet the application still doesn’t do what its users actually need, or what the business who produced it wants. Or, conversely, a usability test will identify myriad usability problems, all of which would be terrible if what the user was trying to accomplish was a simple purpose, but which are much less important, or irrelevant, when their complex purpose is correctly understood.

Let’s see an example of the difference in the predictability of ‘using’ a thing vs. its true fitness for purpose for a machine with a simple purpose, such as a car, vs. an application with a complex purpose, such as accounting software.

The Four Levels of Application Quality & Testing

In conclusion, we can summarize all this by proposing four levels of application quality, all of which have to be tested differently, against different criteria that are defined by critically important work that takes place at different steps in the software planning, design, and implementation process. Note that #4 is usually omitted, or taken for granted, or done badly.

1. Execution: no bugs, executes without errors, based on code-level test procedures. Testing often automated.
2. Functional: everything works as per the design, faithful to the design. Requires human inspection – this is where “acceptance testing” usually lives, and where contractual responsibility ends for any third-parties developing the application.
3. Usability: easy to understand, easy to control, quick to use , based on interacting with the UI to accomplish the tasks that the application is designed to accomplish (defined in the conceptual design or ‘requirements’)
4. Fitness for Purpose: enables users to accomplish each of their relevant aims easily and effectively while generating the benefits the business wants from it. These criteria are usually defined as part of the strategic plan for the application. This includes branding criteria and other non-functional characteristics that should be included as important components in the experience of the application.

New Metaphors for Usability

If we ask the question, why does the usability field suffer from this blindness about complex purposes? One partial answer is that the discipline of usability emerged historically from industrial-design testing, and industrial design is the design of machines and  tools, most of which are instruments with simple purposes. We see the same blindness at work when software developers are referred to as “engineers”. Website design is, by comparison, the design of systems of informing and communicating. The right metaphors are ones based on communication between sentient beings, albeit one side is much stupider than the other. Some initial suggestions below:

• “A website is a form of communication.”
• “Usability is communication quality.”
• “A user is a participant in a dialog.”
• “A website is a virtualized person, group, or company.”