6: Make It Happen: Build for Extreme Usability and Business Evolution
Usability is much more of a critical component of customer relationships than in any other channel. Any user, even if he is new to a site, expects to be able to complete his task rapidly, with a minimums of effort, and will not hesitate to leave one site for another if he encounters any difficulty. As a result, any usability problem will have an immediate negative effect on the business success metrics like retention and browser-to-buyer conversion.
In the software industry, usability was often a last step, when it was performed at all: software was tested to make sure it worked properly and that the user could figure out how to work it properly. If there was any issue, it was noted for inclusion in the instruction manual.
As companies go online, user experience has become a major source of competitive advantage: many sites offer essentially the same merchandise at roughly the same prices, so the reason customers prefer one shop to another is often the experience of using the site, a critical component of which is usability - the mere ability to complete a task.
A site's usability can be witnessed in its click-stream data: the point in a process at which users abandon a task without completing it. Large retailers online do extensive testing of their interfaces, monitoring results in real time, to determine whether a change adds to or detracts from the willingness of users to click through to the next screen in the flow and, ultimately, to the end of the transaction.
Usability doesn't occur accidentally, but s part of the "culture of development" of a company. The firm must understand and accept that user experience impacts their bottom line and, as such, to take a user-centric approach to their online operations.
(EN: this sounds very familiar to me - it's similar to the way that brick-and-mortar retail underwent a change to become more service-oriented rather than merely a warehouse with a checkout counter, and come to realize that the many "extraneous" elements of the customer experience, from the moment they entered the shop to the moment they left, had a strong influence over buyer behavior and, notably, customer loyalty.)
For online retail, translating usability into dollars is fairly straightforward: if design changes result in a 5% increase in transactions completed, that amount can be said to resulted from usability. For other products, it can be more difficult to measure whether the customer's interaction with a support Web site had anything to do with a decision made, at a later date, to buy an upgrade or additional software from the same provider.
Aside of revenue, cost-savings can be assessed: if the number of calls to a support hotline decreases by 25% after a manual is placed online (or existing documentation of the product is improved), then it is likewise reasonable to attribute the cost savings to the usability of the site.
(EN: Even this is short-term and superficial. The benefit of experience and usability are more long-term, and seeking to make short-term improvements may in fact be harmful. A change that causes a 10% increase in sales for a given item may cause a decrease in the number of repeat customers if they felt they were herded into buying an item that didn't suit their needs.)
Customer-Centrism Is a Cultural Mindset
The mindset of software developers has long been that if their software executed correctly, then their job was done - and it was up to the user to figure out how to interact with the software in the proper way. (EN: the popular acronym of the day was "RTFM"). This mindset was largely carried over into Internet and new media channels.
On the level of the firm, usability is often recognized. Amazon.com largely led the way: they gained first-mover advantage by being the only online bookstore at a time - but when competition from more established retailers moved in, they maintained a lead by making their site the easiest and most user-friendly option. They are famous for their customer research, and being user-centric is at the core of their company's culture. This has played out repeatedly on the Web, where small and unknown firms have risen to dominance, overtaking more established first, by virtue of their focus on user experience.
Some random tips:
- Ensure that the technology team understands that usability, not merely functionality, must be considered as a part of software requirements
- Validate that the design meets usability criteria for every audience segment (inexperienced users as well as experienced ones, international as well as domestic, etc.)
- Design interfaces to collect and report on interactions in real-time
- Do extensive testing with actual users, and require the team to observe usability testing (in real time or later on tape)
- Create a "learning culture" that maintains an organizational memory of design techniques that have fared well or poorly in testing as well as in practice
- Tie compensation and rewards not to the completion of tasks, but to the impact that they have on the financial success of the site.
Developing software that serves as an interface to the customer, and that has a direct impact on the company's top-line revenues and bottom-line profits, has resulted in intense concern about technology from outside the IT department. Most Web projects involve a cross-disciplinary work team that include marketing, design, technology, business operations, legal, and other disciplines.
As an aside, the author mentions "experience" as a notion that must be redefined. An individual who has twenty years experience designing bad systems is less important that another who has four of five years of working on systems that have been successful.
Most often, Web site operations are funded and controlled by the marketing department, as the site is a channel through which the firm interacts with customers and prospects. Marketing will either want to "own" the user experience, or will ride herd on the IT department to make sure it's to their liking. However, the agenda to both of these departments may be contrary to user experience.
Wherever the task is assigned, the owner of the user experience must balance the goals of the business against the needs and desires of the end users, and must be empowered to negotiate for the user against competing priorities.
Currently, there is no formal training path for user experience professionals, and they come from a wide array of backgrounds (design, IT, marketing, engineering, quality assurance) and have an even wider array of educational credentials (business, engineering, psychology, anthropology, communications, computer science, etc.)
(EN: Again, this is succotash, but based on personal experience, I can do no better. UX professionals are a motley group, and I haven't noticed that any specific education/experience profile seems to excel at the job more than others, though many of them seem to come from design or IT out of their proximity to the task.)
Usability Metrics Will Be Tied to Business Metrics
The author offers anecdotal evidence that usability tests are more effective if the results can be translated in terms of business-success metrics. You will get less consideration and support from executives if you indicate that users "seemed confused by this", but they will pay close attention if you can indicate that it prevents users form completing a purchase.
(EN: Again, I balk at short-term metrics. You can boost immediate sales if you remove some of the "confusing details" from a flow - but will have greater long-term loyalty, hence income, if you seek to ensure customers were able to make an informed buying decision.)
The author suggests using both qualitative and quantitative criteria in usability testing. Measuring the number of seconds it takes to complete the task is a good measure of usability - but it's as important, if not more, to ask if the user "feels" that the task took a reasonable amount of time or effort.
Evolutionary Software-Development Process
The author contrasts the "waterfall" method of software development (where a long time is spend developing requirements before coding even begins) versus the more current "evolutionary" method of development (bang it out, make repairs and improvements after release, all very rapidly), with an obvious preference for the latter.
(EN: I can't agree. While Microsoft has made a standard practice of releasing buggy software followed by multiple patches, it's has frustrated and angered their customers. Especially for user experience, taking the time to do things "right" is the better path: the user who comes to your site and has a problem will not come back later to see if you've fixed it - he's already lost to a competitor.)
The author suggests that the team can release software that provides the "major value" of the site to customers - i.e., the most critical functional areas - and then further evolve and add features and functionality in later phases to the site.
(EN: This may be more reasonable - but even so, it will only work if you are leading the field. If a competing site already offers the 'extra" features you intend to roll out later, then your site seems inferior by comparison and, again, the customer who is disappointed in your lack of functionality will be lost to a competitor, and not return later to see that you have added it in.)
(EN: One last barb for the "evolutionary" approach is that it results in a site that is constantly changing. While it was once thought that users were impressed that a site seemed to be always changing and adding new features, it's since been found that repeat users are more frustrated than delighted when a site changes.)
The Changing Role of Usability
In comparing usability for the Web with traditional software usability testing, its role has changed: It used to be conducted in the R&D lab to address tactical concerns (find bugs) when a software product was 99% completed. It is presently conducted by marketing to identify strategic opportunities and is part of the product design process (and sometime, is done beforehand).
The author concedes that usability testing hasn't widely been adopted, for a handful of reasons:. Primarily, it is expensive and time-consuming to prepare and execute a test scenario and analyze the results; because it's generally done with a small number of test subjects, its statistical level of confidence is rather low; and if usability is easily botched, especially when it is handled by amateurs or those with an agenda.
(EN: As such, usability is often best outsourced to a company that has a lot of experience at it, and has no agenda in "proving" or "disproving" the prototype is usable.)
What Can, and Cannot, Slide
When a project moves to execution and developers begin coding out the solution, they will often seek to negotiate for compromises to the design, generally to reduce time and effort in the development task, eliminate redundancies, improve performance, or a variety of other reasons. It should be possible to let some things "slide" without degrading the quality of the user experience, but others must be rigorously defended.
The cosmetics of a site do not have to be perfect. If a button is fiver pixels lower on the screen than intended, or an image seems a bit grainy, that is not a show-stopper. (EN: but neither are cosmetics to be entirely dismissed - for example, if a product image shows the wrong color, that could be a serious problem for some customers, and it needs to be addressed before launch.)
There is also some room for compromise on the degree of usability. The site and its core functions must operate and be accessible, but it need not be optimal. It stands to note that "optimal" is a level of perfection that operators constantly seek to achieve - there will always be areas where improvement is necessary, and minor difficulties will surface no matter how much effort is put into the design.
The author enumerates some of the reasons firms deliver untested solutions to the market:
- The culture of the firm doesn't value it - which is common in engineering-driven firms (if it works, it's good enough, and it's the user's problem to figure out how to work it)
- Narcissism - the project sponsor or key team members are fond of their own opinions, and do not see the need to get input from customers.
- Arbitrary Deadlines - Marketing sets an arbitrary deadline for a project to "launch" that doesn't provide sufficient time for testing
- Operational efficiency - Similar to arbitrary deadlines, project managers seek to make the process more cost/time efficient by eliminating usability as an "unnecessary" step, either from the onset or when the project falls behind schedule
The author also lists some of the reasons firms usability is done poorly:
- Expert Users - the site is tested internally, or with a group of individuals who are familiar with the company and its products
- Assumptions - The testing presumes that users will want to perform a given task at all
- Test Flaws - There are numerous problems, such as testing the wrong users, asking leading questions, assuming a singular path to success, etc.
- Analysis flaws - Results may be misinterpreted, either skewed by preconceptions or merely making inaccurate guesses as to the reasons for observed actions
- Narcissism - Individuals within the business blatantly dismiss or disregard test results, asssuming their original ideas to be perfect and the users to be flawed (stupid)
Launching and Ongoing Refinement of the User Experience
Once a site is launched, the task continues: you observe user interactions and can observe, within a matter of hours, whether the site is successful and what elements need improvement.
The key tool is clickstream analysis: observing the path that users follow through the site, from the point where they arrive to the point where they depart, with or without completing an expected task. Particular attention is to be given to the "exit" pages, especially when a user leaves the site without finishing a task, as this identifies a page in which users are likely to have encountered difficulties that led them to leave.
(EN: Assumption is both a prerequisite and fatal flaw of clickstream analysis. The analysis is based on the assumption that the user seeks to perform a specific task and follow a specific task to do it, and can give "false negative" results if these assumptions are incorrect. As such, it is to be used with care.)
The Future of Usability
The author forecasts that usability will play a critical role in the development of user experiences. Given that there are multiple options available on the Internet for users to accomplish their goals, they will choose the best among them, and this is often based on the level of usability.
(EN: There's some argument to be made over whether usability or usefulness is of chief importance. Some usability experts have been amazed that ugly and awkward sites have performed well, and even been preferred by users, over sites that were easier to use, but less useful in accomplishing a task. As such, I take the perspective that usefulness must come first, with usability close at its heels.)
The continued evolution of the channels make usability an ongoing challenge: we must constantly venture into unknown territory, as each new device has different capabilities and different patterns of user interaction. The principles and methodologies must be tested and reinvented for each new channel.
Especially for mobile media, context plays a greater role. Testing a mobile application in a quiet laboratory setting will yield different results that the user "in the wild" who accesses the application while walking through a busy airport terminal. This will require adaptation to get accurate results for usability.
The author also notes a trend toward smaller sample sizes: companies are doing usability tests with a very small number of users, which itself already leads to a decrease in the reliability of results.
It's predicted that usability testing will become a more specialized discipline, such that it will be less feasible for a firm to have an in-house staff, and more likely firms will turn to independent laboratories for usability testing, and may engage different firms that are specialized to a given medium and facet of usability,