In This Issue:
- Consumer Acceptance of Speech Recognition
- VocaLabs Wins "Best of Show"
- Smoother Surveys, Better Results
- VocaLabs Tops 25,000 Panelists
By Peter Leppik
We recently completed a study on consumer acceptance of speech recognition technology, and I will be presenting the results at the upcoming V-World conference in San Francisco.
For those who want a sneak peek at the results, we reached some interesting conclusions:
* Most consumers don't have strong opinions about speech recognition technology.
* After trying a speech application, callers viewed it as superior to tones in every quality we asked about.
* Callers liked the test application, but wanted to be able to do more: more information, more customized.
This is consistent with what we've seen in surveys we've done: speech recognition technology, if implemented properly, is far more popular with callers than tone-based applications.
But when building a speech recognition application, there are also more things which can go wrong. Implementation and design have to be done carefully, to ensure a quality result.
By Jessica Ament
VocaLabs won "Best of Show" at the recent AVIOS/Spring SpeechTEK conference in San Jose.
AVIOS/SpeechTEK Best of Show
The AVIOS board cited our service as filling an important role in the industry, as well as our unique format for presenting survey results to clients, and our ability to produce concrete usability and caller preference data with a rapid turnaround.
We're extremely proud of the recognition we received at this conference, especially since AVIOS is one of the oldest professional organizations devoted to speech and related technologies, and has a history of focusing on real-world results, as opposed to slideware and PowerPoint.
As speech recognition technology becomes more common in customer care operations, the service VocaLabs provides becomes even more critical. Speech recognition has moved well beyond the "science project" stage, and directly impacts a company's relationships with its customers.
VocaLabs ensures that your customers are receiving the best, most cost-effective service possible.
By Rick Rappe
VocaLabs has been making a number of service improvements recently, to make the survey process as smooth as possible, and make the data as easy to interpret as we can.
The first of these improvements was announced back in January, when we rolled out on-line real-time results. This allows our clients to view a survey as it progresses, though a private web page.
Real-time results aren't half-measures, either. The results are provided in the same format, with the same tools, and the same data, as the final survey results. You can listen to call recordings, view graphs, and download the results to a spreadsheet as panelists complete their surveys.
Just this week, we announced the availability of an on-line survey prototyping tool. This allows our clients to view a prototype survey through a private web page exactly as panelists will see it, smoothing the process of defining a survey.
For those with enough volume to be writing their own XML survey definitions, the on-line prototyping also provides a quick way to ensure a valid XML file.
Those of you who saw us at AVIOS, or met with us in the past few weeks, may have noticed that our deliverable format has also been considerably enhanced. We've added the ability to graph results from subsets of panelists, and extract subsets of panelists based on multiple criteria.
For example, if you need to examine just those panelists who were assigned the task of changing a PIN number, but who couldn't finish the task, three clicks and you're there.
The enhanced deliverable format hasn't been formally introduced yet, but we expect to announce it soon. In the meanwhile, if you want a sample deliverable in our new format, contact me, and I'll send one to you right away.
By Jessica Ament
VocaLabs recently topped 25,000 panelists (as of today, the exact number is 26,084).
This allows us to offer larger surveys and more precise demographics. Larger surveys mean more precise results.
For example, many surveys performed by other companies only have around 100 participants. With 100 participants, the margin of error is at least 20 percentage points. This means you can distinguish between 70% and 90% satisfaction, but no better.
We generally recommend not doing surveys with fewer than 500 participants. At 500 participants, the margin of error is nine points. If you go up to 2,500 participants, the margin of error drops to four points, with the same degree of statistical confidence.
Just as important, we can now perform larger surveys with a more specific demographic profile. If you have a service operation which only serves a particular region of the country or a particular demographic group, we have the ability to be more specific about the survey participants.
If you have a particular question about a specific survey size and demographic profile, please contact your VocaLabs salesperson for more details.