System Integration: Keys to Finding the Right Design Perspective

Imagine your home was built in several different eras and styles: The entry is ornate Victorian, the living room is angular 50’s modern, and the bedrooms are done in a country style. As a result, the house is aging and unattractive, but functional and structurally sound. As a first step towards a “face lift”, you added a new guest room last year. The new room is so fabulous it was included in a home magazine (but it’s in the back, and most people never see it). What’s the best way to make noticeable improvement to your home, if you’re on a budget? Obviously, you’d make aesthetic improvements throughout, focusing on the most used rooms of the house—but avoiding big structural changes. Believe it or not, we’re in the same situation as owners or designers of over-the-phone services. Our users enter through touchtone (often recorded in several voices), and may have speech somewhere in their interaction. They receive inconsistent information in different applications, and are sometimes asked to log-in more than once. The end-to-end user experience is often disjointed, unbranded and less than optimal—despite the attention paid to the new speech system last year. What can we do? The design of voice user interfaces has come a long way in the past four years: A process exists which focuses on designing so that a system is easy to use, useful and compelling for the end-users, and is on its way to being embraced by the industry. Conversational, persona-based design is proven to be easier for callers to use, and more satisfying. Principles, approaches and templates exist which allow best practices for voice user experience to be taught. Business people assert that an improved user experience is one of the great value-adds of a voice interaction (as attendees of this year’s SpeechTEK consistently reinforced in their sessions). While we have our work cut out for us, as designers of speech systems—we have good tools in our toolkit. But that’s not enough. In the current environment, companies (like our homeowner above) need overall improvement without breaking the bank by starting over. For speech to reach its potential, we need to think of ourselves as designers of end-to-end services--which means thinking about speech in the context of other modes, like touchtone, Web and human operators. That’s the way callers encounter speech recognition: as one part of an overall usage experience. They may call a speech-enabled customer service application that transfers them into a financial institution’s legacy touchtone IVR. They may reach a speech-enabled FliFo application, after pushing a couple of touchtones. Or they may have a watch list, which was set up on the Web and used while driving in the car. Businesses will realize the value of speech in the same way: In order to become more than a “neat, next-gen technology” tried out in the call center, speech must have a unique role that’s integral to the planning and execution of customer service programs and brand marketing across the enterprise. So, as designers of over-the-phone voice interactions, we need to think about more than just the speech application Three philosophies that can change everything. 1. Stop being religious about speech. Yes, we all drank the kool-aid and saw the research results: speech is preferred by users to other modes, for all kinds of reasons. And yet, no one’s throwing out their existing touchtone systems. Many companies actually spending more on the Web, despite the fact that about 70% of company contacts, on average, occur on the phone (Society of Consumer Affairs Professionals Research, 1999), and their customers are more mobile than ever. Today, the user is truly platform agnostic. In their effort to get things done, most consumers use more than one mode—and sometimes more than one brand. As a result, companies understand that they need to increase customer satisfaction and loyalty across all contact points. But surprisingly, most companies manage their telephone infrastructures in multiple, separate groups and their Web infrastructure in its own, siloed organization. The result? Huge variability in service quality and brand communication. At my company we’re currently reviewing one company’s system in which it’s possible to be connected to a touchtone service and then be transferred through two other systems (one using speech recognition) to complete a single task set. Each of these systems has different voices, prompting styles and “universal” frameworks (help/error recovery and universals commands). In this case, a well-designed speech system might not necessarily be a value-add—it might just be another source of confusion. Even worse, inaccuracies can exist, causing the information consumers get on the Web to be different from the information they get on the phone. In recent research, this discontinuity undermined the caller’s confidence in both systems: How do I know the company’s got it right? Bottom line: other contact modes are not going to go away, and voice will only be recognized as a superior form of interaction if it fits seamlessly into the context in which it is initially asked to reside. 2. Think in context about what speech can add for the user or business. As designers, we know that it’s vital to consider the user context to create a useful, usable interaction: who are they and what are they trying to accomplish? Going forward, we need to consider the usage context as well: Companies have multiple channels. Users are almost always exposed to some other channel besides speech—either within or across usage experiences. This understanding should impact your recommendations about the role of speech and its specific design. Recently, a client told us about a newly completed, voice-enabled branch locator application, accessed after a touchtone interaction. Given that the branch locator is a convenience item, and is unrelated to other tasks achievable in the IVR, this new service can’t add its intended value in its current position. No matter how elegant the dialog design might be, the branch locator application has lost its convenience through its placement within the overall IVR structure. We must think through the entire task structure and optimal user experience, and make speech available in the appropriate context, where it adds the most value. 3. Understand that branding is a corporate-wide initiative—and is an intrinsic form of consistency we can offer the caller. Companies seek to offer a consistent image to the caller, across all marketing media: TV, radio and Web. Telephone-based services should be included on this list, but until now they have been largely ignored. Consumers see the difference: automated telephone services have been seen by callers as a cost-cutting measure for the companies that deploy them—touchtone systems are disliked, even when they’re effective, and are seen as a way to “keep people out” (Speech Technology Magazine Research conducted by Michelson & Associates, 2002; and other studies). In contrast, voice enabled services engage the caller and create a sense of cooperation between company and caller—making it possible for callers to collaborate with the brand itself to get something done. Studies have also shown that effective voice interactions can improve feelings about the company and the brand. But this result isn’t automatic. It requires a focus on bringing a brand character to life by creating a “persona definition” early in the process, and instantiating that persona in every aspect of user experience: from dialog design, to prompt wording and recording. Avoid persona at your own risk: Research shows people will automatically make associations, positive or negative, with any voice or prompting style you choose. Why not plan it in advance, to ensure relevance and consistency? Right now, many companies are developing personas as part of the design process for their new speech interactions. However, like the branch locator mentioned above, they’re just one piece of an overall experience. Research shows that impressions are formed in the first seconds of any interaction—so by the time the caller reaches the carefully crafted speech persona, their feelings are already set. To be effective, persona cannot be decided at the application level. It must be identified corporately, along with the brand identity, and implemented consistently across operational teams and applications. To ensure success, consistency of telephone user experience must be managed horizontally—currently, a missing role at most companies. What can I do now? In the current climate, technology purchases are difficult—but it’s a great time to optimize what you’ve got (to ‘redecorate’ in a consistent style, without significant structural changes). Right now you can work with customers to upgrade the overall user experience quite simply and inexpensively, through the following steps: 1.Define a corporate persona: Meet with marketing strategists or the advertising agency to understand what the core values of the brand are. Translate those values into a persona. Remember, you are not trying to create a (fake) person, you are trying to make it possible for the caller to infer the same positive characteristics from the over-the-phone experience that they have of the company and brand overall. This exercise serves the dual purpose of helping redefine the role of the phone for marketing and business executives, while creating a framework for interface consistency. 2.Conduct a usage and usability review of all systems, to identify ways to a) improve usability in the short run, without substantial code changes; and b) ensure cross-system consistency and convenience. Through this exercise you will develop consistent terminology and universal design strategies that extend across systems. 3.Edit (rewrite) and record prompts across all systems—touchtone and speech—to improve both usability and branding—creating one seamless, branded, easy-to-use service set. Right now, it’s also a good time to plan for the future. Conduct any additional research you need to understand the role of the phone in the user’s life, and determine how speech can contribute optimally to caller satisfaction and business goals. An independent third party is often a good source of this evaluation, as they are platform/software agnostic. Make sure your company has a holistic evolutionary plan for bringing the phone and the Web together, to provide superior, consistent, branded customer service across all contact channels. Not only will this improve service today, you’ll be ready for multimodality as the technology matures. Bottom line: A consistent, branded, end-to-end experience is more effective for implementing companies and their callers. While you may be designing a voice interaction, it surely won’t exist in a vacuum. Callers interact across modes or channels, and good speech design must reflect that.
Melissa Dougherty and Wally Brill are both principals at Voice Partners LLC. They can be reached at melissa@voicepartners.com and wally@voicepartners.com.
SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues