Speech Technology Magazine

 

Data Is Indispensable— Except When It’s Not

Sometimes the right business decision is so obvious you don't need data to prove it
By Jenni McKienzie - Posted Jun 14, 2016
Page1 of 1
Bookmark and Share

I love data. Data is good. I’ve had colleagues who called me the Excel Queen. I’ve had discussions with people about favorite Excel functions (VSORT and CONCATENATE can do lots of cool things for you). I am ridiculously happy spending an afternoon with my head buried in data, trying to figure out the story it has to tell. I think there’s not enough data to call upon in voice recognition and speech design. I wish there were more pure research being done. It would be wonderful to have that kind of data to back up recommendations to clients.

But sometimes it’s OK not to have data, and it’s OK to base decisions on something other than data.

There’s a lot to be said for common sense and experience. An expert opinion isn’t pulled out of the air, but rather based on loads of experience that often has no hard data behind it. “We need data” is sometimes used as an excuse for not using common sense.

A question came my way recently: Did I have any data measuring the impact of employing a professional voice artist versus using internal or non-professional voices? My initial reaction was no, I did not have that data—but seriously, why would you need it? (And there may be such data out there, but I didn’t see why I should spend any time at all trying to track it down.)

But the question made me think about the underlying logic. Why isn’t data necessary in this case? I came up with a parallel situation. Think about local commercials you see on cable stations. You know the ones I’m talking about; they grate on your nerves and make you roll your eyes and consider changing the channel until they’re over. The acting is so bad that it’s obvious the advertiser got friends and family to help out. What is your impression of this company?

I find it interesting that the question came to me as voice artist rather than voice actor. Because I think the acting part is important. Not everybody does it well (even if they think they do). I don’t even like the phrase voice talent as much because talent somewhat implies it’s either something you have or you don’t. There’s a lot of skill in acting. Like many things in life, as you learn and internalize skills, they become things you can do automatically without thinking of the learning that went into it in the first place, which is when it begins to look like talent. But that doesn’t discount there was learning somewhere along the line.

Image matters. There is hard data around that. Impressions matter. What’s the impact of a bad haircut? How much effect does ill-fitting or inappropriate clothing have in an interview or a first date? Do you really need hard data to know these things have an impact?

If I call up your company and the inflection of the prompts is off, it could lead to turn-taking issues and error conditions, or maybe just my feeling that the experience wasn’t up to the level it should have been. Will that be enough to make me leave your company? Maybe, maybe not. It depends on how bad the experience was and what other things have happened. Maybe it’s enough to make me tell a friend who asks about my day that I had a bad experience. The little things add up.

To go back to the original question about voice actors, it should be enough to answer with this: A non-professional lacks the skill set of a professional (thus the term “non-professional”); the result, then, will be suboptimal and reflect poorly on your company, brand, and image, to some degree. (Just how much, of course, depends on the gap in the skill set.)

That should be enough. Applying logic and rational thinking should tell you all you need to make this decision. If A is better than B, and there’s not a big cost difference between the two, do you really need to know how much A is better than B?

Trust your common sense. Trust the experience of your design team. Trust that they’ve internalized learning and skills to develop a talent for doing what they do. Don’t ask them to go dig up the data and provide you with their entire learning base to move forward with recommendations. Data is good, but it’s not everything.


Jenni McKienzie has worked as a consultant with SpeechUsability, a VUI design consulting firm, since 2013. Previously she held positions at Travelocity and Intervoice. She is also a founding board member of the Association for Voice Interaction Design.

Page1 of 1