Why Your Next Survey Might Have A Microphone

David Ferris
4 min readDec 15, 2020


Surveys are broken. We loathe them when they land in our inbox, and if we do manage to complete their long lists of tenuous multiple choice options, most of us feel like our submissions disappear into an empty abyss. This general feeling is called survey fatigue, and it has eroded response quality and completion rates for the past 2 decades. The PEW Research Center says survey response rates have fallen from 36% to less than 10% from 1997 to 2020, and many report a decrease in response quality as well.

So why do we still use online surveys? Simply put, they’re incredibly inexpensive. At almost no cost, a researcher or marketer can design a survey and email it to thousands of users. Even professional panels (where people are paid to complete surveys) cost only a few dollars per response. This cost-effectiveness combined with flexibility and ease of online distribution have caused surveys to grow tenfold in the past decade.

This growth has been accompanied by a corresponding increase in survey fraud. Professional panels financially reward people for completing surveys, an incentive that rewards speed over quality. At best, responses are erroneous or prone to bias. At worst, they are downright fraudulent. For scientists and researchers, sincerity is paramount and this fraud represents a serious problem.

Survey programmers have gotten clever over the years in an attempt to combat survey fraud and fatigue. Their techniques include increasing sample size, asking the same question twice to verify consistency, and discarding responses completed unreasonably quickly. Unfortunately, these attempts are limited in their effectiveness because they do not address the underlying user experience flaws that deteriorate survey feedback. Since these flaws may be fatal for the survey as we know it, it is essential that we explore better, more human ways of performing research at scale.

Improving richness and accessibility with voice

Enter Phonic. The recent Y Combinator backed startup has launched a survey platform designed to collect more human feedback by collecting spoken responses instead of conventional typed ones.

So why voice? Oral communication is tremendously efficient. We speak aloud at about 150 words per minute compared to just 40 when typing, so it’s no surprise that people provide three times longer and two times more descriptive responses when answering with voice instead of text. Beyond this, voice contains a tonal component not present in the text alone. This tone, in addition to other non-verbal cues, represents more than 90% of human communication.

The analysis of tone of voice is a large component of the field of qualitative research, whereby researchers conduct interviews, focus groups, and otherwise analyze human data. This approach lets researchers prioritize certain feedback, dig deep on specific insights to weave a compelling story.

By contrast, the purpose of a survey is to get the high level takeaways without having to read every individual response. More survey data doesn’t typically provide profound new insights, just greater statistical confidence.

Sitting squarely between the quantitative world of surveys and the qualitative world of interviews, voice-based feedback doesn’t fit this dichotomy. This is a paradigm which Phonic is happy to break, however, since traditional methods are simply a tradeoff between data quality and quantity. This ability to analyze voice feedback quantitatively at scale as well as in a more granular, qualitative manner is a legitimate superpower.

What if I don’t want to speak into a microphone?

Sometimes we’d rather answer with text instead of voice. Maybe we’re on a crowded bus, library or for some other reason can’t or don’t want to speak aloud. Fortunately, most Phonic surveys allow for a text box backup when the respondent would rather not speak aloud. Phonic’s goal is to get better insights with lower friction, and they certainly don’t want to exclude those unwilling or unable to answer with voice.

What’s next for the survey?

What’s richer and more descriptive than audio? Video of course! Some people might be aghast at the idea of recording and sending a video of themselves as part of a survey, but it’s important to remember that there are a huge variety of survey use cases. Video might not be appropriate when collecting feedback from people we don’t know, but surveys are used inside of teams, organizations and for employee feedback. Phonic also uses techniques to anonymize voice and video data that may be personally identifying.

Find out more about Phonic here: https://phonic.ai



David Ferris

Engineer, runner, and startup founder | Check out Phonic (www.phonic.ai)