By Jerrold Prothero, Ph.D.
In the gold-rush days of the web, it was truly remarkable how little money was spent by start-ups on usability testing. What was even more surprising is how few investors insisted that usability be built into the development process. Had investors demanded a vigorous usability feedback loop and monitored the reports, they would have very likely been alerted to the fact that the product was in trouble.
Investing millions in a failed venture is no joke. Nonetheless, there is a little witticism in the usability community that says the typical product development life-cycle consists of the following steps: design it; build it; test it; find out it doesn’t work; and ship it, because there isn’t time to do anything else. This life-cycle produces an enormous number of expensive customer service calls, in addition to losing customers entirely. The saddest truth of all may be that the lack of early and consistent usability testing may have been the root cause for many dot.com meltdowns.
While we all understand that software engineering is concerned with a product’s functional requirements, we are often unaware that “usability engineering” is concerned with the customer requirements for a product. Usability engineering starts from the customer’s perspective, proceeds to the design of an interface to meet the customer’s needs, and then addresses the functionality needed to support that interface design. The interface is how customers interact with the product: from the customer’s point-of-view, the interface is the product. Ease of interaction is the critical factor in the success of new products. A good interface can make up for bad functionality; but good functionality can not make up for a bad interface.
A common (and damning) error in web site development is to focus primarily on the site’s appearance. The telephone book is not in every home and office because of the quality of its graphical design. It is there because it efficiently satisfies a need. In my years of usability testing for web sites, I have never had a participant pound their fist on the table and shout “this site needs better graphical design!” Complaints about confusion or difficulty using a site, on the other hand, are continuous. In one case, a polite comment was: “I still don’t know what motivated this company to exist”.
Currently, development decision are often made in meeting on the basis of competitive hand-waving, when they should be made on the basis of usability data. The interface can and must be designed, prototyped and customer tested from the very beginning of the development process. This includes inexpensive customer testing of early paper prototypes, which can provide valuable feedback on general task and layout requirements.
Developers and investors have a lot in common; they are often too close to the product to be able to see usability problems from the customer’s point-of-view. Both need to see and hear how potential customers are reacting to the early build-out of a product. When monitoring an investment, would it be better to be routinely reviewing usability test results from an objective 3rd party, or to look over another notebook of “tasks completed” at an “everything’s great” presentation? Which of these brings the venture closer to a profitable product?
My corollary of Murphy’s Law is that anything that hasn’t been tested, doesn’t work. Everyone knows this for software. If you haven’t tested software, it will fail. Exactly the same thing is true for interfaces. If they haven’t been tested, they won’t work. The difference is, when software fails, it is usually obvious. Something crashes. When there is a problem with the interface, it is much more subtle. You simply annoy potential customers, who disappear. Interface usability problems are as insidious as high blood pressure: they kill you, but they kill you quietly.
A few facts on the value of usability testing and good interfaces:
* Just one cycle of interface analysis and redesign produces on average a 50% improvement in end-user performance.
* Usability engineering reduces the product life-cycle by 33-50%, by providing up-front information about what functionality is needed and avoiding major re-designs.
* Catching interface design problems early in development through interface analysis is 40-100 times less expensive than fixing them after initial product release.
* The cost of a single support call has been estimated at $12-$250, depending on the application.
* Two-thirds of software projects significantly over-run budget and schedule.
These few simple facts can tell an investor an enormous amount about how valuable usability testing could be to the overall success of the venture.
Too many rounds of funding have given up the ghost through mistakes that we cannot afford to repeat. It is my suggestion that the investor community build usability testing into their funding requirements for the development process.
These 5 simple points are a good start towards securing your investment:
1. Know who your customers are; profile them early.
2. Analyze the tasks your customers will need to perform.
3. Design the product from the interface, through to the functionality. Not the other way around. Again, remember that from the customer’s point-of-view, the interface is the product.
4. Involve usability testing from the very beginning. Early on, test paper prototypes or Powerpoint mock-ups.
5. As the product develops, continue user testing in parallel with software testing.
The above five points are a beginning. The general idea is to build the customer into your development process. In the end, your customers will decide the fate of your product (and your investment), so let them help you design it. Go looking for reality, before reality finds you.
Who is Jerrold Prothero?
Dr. Jerrold Prothero one of a small number of people with a Ph.D. in human-computer interfaces, awarded through the Human Interface Technology Lab at the University of Washington. He is president of Hypercerulean, Inc., a Seattle-based consultancy specializing in the human factors of e-commerce and emerging technologies. He can be contacted at firstname.lastname@example.org.