Web Sites Ethiopia

The intent of this chapter has been to show how usability can be factored in throughout the design process.  Usability needs to be part of every task in the design and production process.  In this book we’ll be walking through every task and explaining the usability aspects of each.

However, as we do so, we’re covering a number of methods that won’t be used in every project.  How do you choose among them?  How do you decide which methods are relevant to a given project?  This will be answered as we address each method in the book.  In this section, we introduce the framework for making these decisions.

Conceptually, it all boils down to a cost analysis.  The argument for usability will pivot on what value the method provides and at what cost.  Methods may achieve value in the form of improved profitability, marketing, employee retention, and humanitarian benefits. 

Value is limited by various risk factors.  How reliable is a method?  How confident can we be in the design guidelines the method suggests?  In addition, cost is not just the amount of time and materials involved in performing the method, but also the cost of training and the impact on the development schedule.

Method Selection Criteria
Table 1.1 shows some of the primary criteria for selecting among various usability methods.  These methods are just some of the more representative methods for each stage. 

Each method can be implemented in a number of ways, so instead of listing exploit times, we’ve simply distinguished between long and short time frames; the detailed information in the following chapters should help in pinning down more exact estimates.  For another point of view of view on comparing methods, see Olson and Moran (1995).

Required Tasks
Certain tasks need to be performed at some level to matter how you develop a web site.  All sites, for example, require you to develop the information architecture, one day or another, whether it’s deliberately and thoughtfully or not. 

You can choose to skip the step of quality assurance, but skipping it will practically guarantee a malfunctioning product.  Similarly, you can ignore user complaints and avoid tracking user problems, but it’s highly unproductive to do so, especially as it cost relatively little (although the cost might be high if your design is awful, but that’s all the more reason to take care of it).

Time to Perform the Method
We haven’t set out exactly how long each usability method takes because this spends on the variables of your project.  However, a short task like reviewing usability checklists can take as little as 10 or 15 minutes. 

A medium-length task such as user testing typically takes anywhere from a few hours to a couple of weeks.  A long task such as running field observations or user surveys can be completed in a couple of days up to several months.

Costs
A variety of costs can apply, although most techniques, marked in Table 1-2 by dashes, can be easily performed with the equipment sitting around a typical office:  paper, pencil, email, and computers with standard software suites. 

A common cost is recruiting and paying users, which often becomes more expensive as you choose a more specific and more narrow, hard-to-find target population.  In the case of surveys, you may need to pay for mailing lists as well as the cost of mailing.  For quality assurance, you’ll need to reproduce and maintain all the target computing platforms, with the associated costs.

Learning Time
How long does it take to learn how to perform each method?  Most methods are relatively straightforward, and learning the useful minimum can be achieved in a matter of hours.  Practice is necessary in every case, though, so the first time you try a new method, plan extra time to run through a few practice trails to get out the kinks.

Confidence Level
Each method will suggest changes to your design.  How confident can you be in those suggested design changes?  How valid is each method?  You don’t want to spend a lot of money making design changes that you can’t rely on.

Probably the least reliable method is to rely in intuition, although an experienced, designer’s intuition may be somewhat better than average.  For the best part, you can have greater confidence in methods that obtain information from actual users, and then greater confidence still the more users you deal with (in some methods, statistical tests may be able to give you a much better measure of confidence level).

Low-confidence methods do not necessarily give you bad information, but you must weight the risks against the costs of other, more reliable methods.  In almost every case, any of these usability methods will help you create a better design, but all of them will occasionally provide misleading information that can undermine the design to some degree.

Impact on Final Design
In general, methods applied later in the process will have less impact on you final design than methods applied earlier.  It’s usually too costly to revisit old design decisions or make radical changes in design as you get further along in a project.

How to Deicide Which Methods to Use
You want fast, low-cost, easy-to-learn, high-confidence, and high-impact methods.  For each method ask yourself:  To what extent is this going to save you redesign time later in the process?  To what extent is this going to save money on the product and save money and time for the user?  Does that justify the cost?

For example, in the requirements stage, should you conduct user interviews or user surveys?  Both will be useful if the budget can handle it, but if your budget is limited, interviews will be better. 

Why?  Because they usually take less time than surveys, and because they provide more detailed insights into what the project requirements should be.  Even though surveys provide input from more users, the types of questions that can be successfully answered are much more limited. 

Should you also perform a competitive analysis?  It doesn’t cost much to do, but on the other hand, the information you receive is often much less useful than that from other types of requirements analysis, and its usability insights are much less reliable.  So, competitive analysis is usually worth doing (because of the low cost), but needs to be supplemented with other information if possible.

Types of Usability
The exact way you conduct your user studies depends on what you want to learn from them.  Three common approaches are as follows:

Mining for problems
The approach to user testing explained in Chapter 12, “Usability Evaluation,” is a common test used to identify problems in a design.  The primary concern of such a test is finding out what problems the user encounters.

Criterion testing
In a criterion test, your concern is, Does this site meet the explicit goals we set for it?  Ideally, at the beginning of the beginning of the project, you’ve defined some usability objectives, things like “most users will be able to complete a product purchase in three minutes or less” or “users will give an average usability rating of 2 or better on a scale of easy (1) to hand (b). 

In this type of test, you focus is on the measurements.  If you’ve met your objectives, you can safely and the design cycle.  However, if you don’t, you’ll need to have recorded the problems in order to keep improving the design.

Comparison testing
In a comparison, you are trying to see which of two or more designs is better.  You’ll want to measure usability in more than one way to guide your decision.  How much time does each task take? 

How many mistakes does the user make?  How do users rate the site?  If one design is better on all these dimensions, your decision is an easy one.  If not, you may be able to identify a compromise design that integrates the best features of each alternative.

Report This Post