im
classic-banner

Web, Mail and IVR Surveys Have Their Own Advantages

In the last edition of MR Perspectives, we explored telephone interviewing as the most common methodology available for marketing research. In this second article of the two-part series, we look at web, mail and interactive voice response methodologies to explain when they are a better choice for conducting marketing research.

Web

With web research, the entire survey is programmed into an interactive web-based script. The technology supports both closed and open-ended questions. The survey is dynamic and can support multiple brands or product types. Thus, only one survey has to be programmed. The code then reads directly from the sample file and either pipes in different text or skips to different questions, depending on the situation.

Invitations containing, at the least, a link to the survey as well as well as login credentials, are sent to everyone in the sample file. If the respondent does not finish the survey, they can complete it later by using the same login credentials to re-access the survey. Generally, one or more reminder invitations are sent to those who do not complete within a specified period of time.

The data set contains answers to each question responded to, as well as minimal information regarding each sample record.

It is best to use web surveys when the targeted respondents access their email regularly and you have access to valid email addresses for the majority of the respondents. It is also best to use web surveys when the sample is coming from a panel, as opposed to a client-supplied list. Also, web surveys have a distinct advantage when it is important to show images or pictures in the survey.

Mail

With mail surveys, the printed survey includes a different version for each brand or product type.�The survey is then mailed to every customer who has had recent contact with the company, typically. After completing the survey, the respondent mails it back to the appropriate place. The data is then hand-keyed or scanned into a database table. The data set contains information regarding those who completed the survey.

It is best to use mail surveys when there is only one or two versions of the survey and you have accurate mailing addresses for the majority of the intended respondents. Mail surveys are particularly useful when the survey solicits sensitive information; respondents are often reluctant to tell a live interviewer how much they earn, or their opinion on a controversial issue such as abortion.

Mail surveys are useful when time is not an issue but low cost is a priority. The mail piece itself (when professionally done) implies a customer care mentality on behalf of the company sending it, regardless of whether the respondent participates.

Interactive voice response (IVR)

With IVR surveys, the entire survey is recorded by a human voice of choice (e.g., male or female).�The technology supports both closed and open-ended questions. All or parts of the survey may be recorded and saved as audio files, such as .WAV files.

The survey is dynamic and can support multiple brands or product types. Thus, only one survey has to be programmed. The code then reads directly from the sample file and either pipes in different text or skips to different questions, depending on the situation. The respondent is taken through the survey one question at a time and is asked to respond by hitting a button on their phone.

The data set contains answers to each question responded to, as well as minimal information regarding each sample record (e.g., whether the number was dialed, if the respondent attempted to take the survey).

It is best to use inbound IVR when:

  • The survey is short (e.g., no more than ten questions).�Typically, unless there are heavy incentives given out, most IVR respondents will stay on the phone no longer than three minutes.

  • The survey questions are easy to understand and are unlikely to confuse the respondent or require additional explanation.

  • Most of the questions utilize the same rating scale because swapping between different scales can often be confusing to the respondent.

  • The �sample size� to �completes needed� ratio is very high. Because the response rate is low, and the completion rate is even lower, a great deal of sample is needed.

  • There are no open-ended questions. While it is possible to record responses to such questions, typically the answers are rather vague and not worth the cost associated with processing the audio files.

  • If there are multiple survey versions, the differences between them are minimal.

  • The survey requires minimal use of quota cell management.

  • The respondents are motivated to complete the survey.

All Content � Copyright 2004-2011. Polaris Marketing Research, Inc. All Rights Reserved.

Send inquiries to newsletter@polarismr.com or call 1-888-816-8700.

ec