Summary


This document provides best practices for setting the survey schedule for Short Message Service and Interactive Voice Response.


Introduction


The Ask tool allows you to customize when and how often surveys are sent to respondents. Specifically, you can set:

  • The days of the week (e.g., Sunday, Monday, Tuesday)

  • Timezone (e.g., UTC)

  • The time of day (e.g., 9am-6pm)

  • The number of times a phone number is contacted (e.g., 2 re-contacts)

  • The time interval between re-contacts

  • When a fallback mode is used, the number of re-contacts before switching modes


Setting the optimal survey schedule is important for achieving a high response rate while not burdening respondents with too many messages at the wrong time.


Guidelines


Below, we provide some general guidelines for setting the survey schedule. The optimal survey schedule most likely varies between each country. However, we also provide some default ideas that you can tailor based on the country.


Do set appropriate days of the week. 

In most countries, a best practice is to send surveys all 7 days of the week to maximize response. In some countries, however, you may have to restrict surveying on a particular day due to religious, cultural, or other considerations, for example, work weeks that run Sunday to Thursday.


Do remember to set the local time zone. 

The default time zone is UTC. You must customize the time zone, otherwise, respondents may not see surveys at the expected times.


Do carefully consider the appropriate start and end times.

As a best practice, we recommend 7am-9pm. In general, a wide time window improves data quality because more people can respond. Cultural considerations may impact decisions regarding survey timing. Legal restrictions can also affect call times.


Do remember that you create one set of times for all days.

Currently, Ask does not support customizing the times based on the day of the week (e.g., start at 10am on Sunday, start at 7am on Monday).


Do set re-contact attempts.

Use re-contact attempts to follow up with people who have not responded to the initial attempts. Re-contact attempts are important for reminding respondents to participate, which improves data quality. We recommend 3 re-contact attempts per mode. If the survey has one mode, you may have up to 4 contacts: one initial contact and 3 re-contacts. If the survey has two modes, you may have up to 8 contacts: in the first mode, one initial contact and 3 re-contacts. The second mode, will have one initial contact and 3 re-contacts.


Do carefully consider the best interval between re-contacts.

As a general practice, we recommend a window of 15 hours between re-contacts. This interval ensures a spread of re-contacts across different times of the day.


Do test different scheduling protocols.

After testing, adjust schedules accordingly to maximize response rates and data quality.


Don't select “all day” for the times unless absolutely necessary.

People are generally unwilling to respond late at night or early in the morning. Sending calls during these times may have legal implications as well.


Don't re-contact people too frequently.

We recommend at least 24-48 hours between contacts. Re-contacting people on the same day may lead to respondent burden.


Don't include too many re-contacts.

Typically, sending more than 6 or 8 calls to a single telephone number has diminishing returns.