VUI Designers Fight a Losing Battle

Article Featured Image

Project plans are beautiful in their simplicity. In the project plans that cross my desk, so many days are allocated for design. There's always a clear expiration date on the design resource...just like milk. Fourteen days for one task, two days for another. It's a matter of allocating resources and juggling budgets. And at some point, the fat lady sings and the VUI designer rides off into the sunset.

A few months later, the system deploys. With a sense of nostalgia, a designer may call in, only to find that the initial-level prompting looks and sounds nothing like the original design.

It's a discouraging fact usually whispered among designers: Often, the most crucial design decisions are made outside of what is generally defined as "the design phase."

As the industry evolves and systems become easier to use and more complex to code due to increasing levels of back-end integration and data usage, there is a lengthening period in any given project between design and deployment. It's the period of platforms and data warehouses, screen pops, Web services, and a plethora of modifiers teamed with the word integration. Acronyms describing incredibly important technical stuff are floated in bulk.

Even with an educated project team committed heart and soul to user-centric design, this is the period where user focus gets lost. If Wizard of Oz or partially developed model usability testing is performed, both taking place earlier in the development cycle, the usability resource who lacks an immediate deadline often retreats into the background during this phase. As a result, by the time the system is ready for user acceptance testing, the word usability is a vague memory in most of the project team's minds.

Which is why prompts such as "Listen carefully, your options have changed" may once again rear their ugly heads. Did the designer or human factors expert spend 20 minutes in the kickoff meeting explaining why this is not usually a good idea? That meeting may just as well have been held when people were still wearing headbands and leg warmers. Due to the placement in the project, these design decisions are often put together in a "defect" list during the quality assurance or user acceptance testing stage, with a note that a prompt needs to be recorded.

Oh! Prompt recording? Pull back in the designer, because don't they do that vocal coaching stuff? Sure, according to the project plan the fat lady sang 62 days ago, but it's only an hour or two.

Except it's not. Because a self-respecting VUI designer is going to look at this mandate and have a bunch of questions and some cautionary tales, and is likely to want to discuss some of these design changes as they pertain to usability and the overall experience.

But often, the ability to have a discussion at this point is difficult. Stakeholders are sending notes and testers are filing defects, and there is a pending roll-out date approaching. A designer who wants to go back to the drawing board and explain why these mandated design ideas are not wise becomes problematic for the whole project team, especially the project manager. It is not a matter of not caring about the end user. But the consequences of a slipping project schedule and a stressed client can be far more immediate than any potential effects on the end user.

And thus, bad design decisions get mandated and coded without much ado. The accountability trail is muddy. And when the system deploys and customers are opting or erroring out, and everyone is asking, "What happened?" it truly isn't clear. Because what really happened was a bunch of well-intentioned and likely talented people got caught up in the details of closing defect tickets, and lost sight of the big picture.

So, perhaps the fat lady shouldn't sing. Because when the system rolls out and for as long as there is a system at all, there always needs to be someone advocating for the user. Whether it's a dedicated person or a rigorous change management process where potential changes are assessed against user impacts matters little. What truly matters is in the effort to deliver something that runs, we don't forget to provide something that works.

Alexandra Auckland is a voice interaction design and tuning consultant at Sotto Voce Consulting. She can be reached at alexandra.auckland@sottovoceconsulting.com.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues
Related Articles

When Bad IVRs Are Good Enough

Users accustomed to substandard systems can't imagine any better.

Usability Testing Connects the Dots

This valuable tool offers unique insights into caller motivations.