Published for Advertising & Marketing Research Professionals by Beta Research Corporation

CAN RESPONSE RATES BY MAIL/TELEPHONE BE IMPROVED?

Over the course of the last few years, response rates to traditional mail surveys have slipped. Several reasons have been offered to help understand this problem.

Most importantly, the variety of competing demands for time has made the completion of a mail survey a more problematic ordeal.

The consumer (as is the businessperson, physician, etc.) nowadays has a host of alternatives with which to deal. The necessity for greater production demands at the office, the rapid growth of e-mail, the barrage of telemarketing, more "junk mail," and family demands are all causal variables pointing to this problem.

Dr. Paul Erdos traced the first "mail survey" in the new world back to 1577 when King Philip II of Spain decided to conduct a census of his new world possessions. In effect this was a mail survey by official courier, since mail service in our sense was not as yet known. In all likelihood, the 1960 Census of the United States provided the major impetus for the acceptance of the mail survey. Since that time, major strides have been taken to bring about a better mail survey.

We have substantially improved our methods for sampling, for questionnaire construction, for printing, and for mailing. Producing a professional mail survey can be readily accomplished.

Beta has previously produced several documents indicating how to raise response rates by mail. Heretofore we have concentrated on the following:

• the outside envelope;
• the pre-alert or advance mailing;
• printing on recycled paper;
• lifting response rates via monetary incentives;
• effectiveness of donations to charities.

Immediately after the disaster of September 11, 2001 we conducted a series of studies that concluded:

• An advance postcard notification, with the name and phone number of a contact person that can verify survey validity, in combination with a response envelope with live postage, increases respondent confidence and results in higher response rates.
• Hand addressing of envelopes does not increase response rates. In fact, when interviewing non-responding recipients by telephone, they seemed to suggest a suspicion of where the package came from. Continued research here is necessary.
• No labels for outgoing #9, #10 or #11 envelopes.
• A Peel 'N Seal (e.g., self-sealing) envelope had no real positive effect.
• Indicate membership in the Better Business Bureau on envelope.
• Use a better grade of paper.
• Use off-color envelope to distinguish the mail-piece from others.
• Mail flat when possible instead of folding in thirds.
• Before mailing, check all lists on databases for proper geographic representation, spelling, abbreviations, lack of specific name, etc.
• Mailing via metered postage is nearly as effective as "live" first-class postage.
• Do not use see-through envelopes and be certain any incentive cannot be seen through the envelope.
• Because business/physician mail is often screened, the recipient must regard the package as "important."
• Print the cover letter on the same paper stock and color as the envelope.
• Cover letters were tested utilizing black, blue, red, green, brown and purple ink for the signature. Unless the signature is designed to match a particular logo color, the standard black or blue signature should be used.
• A two-color, laser-printed cover letter is more effective.
• The cover letter is important. Keep it as short as possible with special emphasis placed on the introduction and the postscript.
• Use easily readable type size for both the cover letter and the questionnaire.
• Hand sign the letter if utilizing a small sample size (fewer than 500) or sending to an extremely difficult population to get to (e.g. CEOs, Thoracic Surgeons).
• When using commemorative postage avoid controversial topics (e.g. planned parenthood, religion).
• Try to avoid open-end questions on page 1.
• "Live" return postage pulls about 3-4% better than BREs.
• We prefer respondents to receive survey packets from Wednesday to Friday; therefore, mailing on Friday or Saturday is preferrable.
• Whereas, several years ago, 80% of all return mail from the first mailing was checked-in within 18 days, it is now necessary to extend this time frame to 24 days. Therefore second mailings should be scheduled accordingly.
• Incentives other than cash almost always yield lower response rates.
• Care must be taken when setting a deadline for return of the questionnaire.
• Priority mail can increase response rates by 4-6%.
• Never falsely claim it will "only take a few moments" to complete the survey if it will take longer.
• Faxing more than a two-page questionnaire tends to have more problems in the transmission of the survey, such as missing pages, difficulty of reading answers, etc.

TELEPHONE
We have found that the introduction should include these messages:

• being a member of the Better Business Bureau;
• we're not a telemarketing company;
• offer our 800 telephone number for them to return the call if at all hesitant.

TELEPHONE/MAIL
Prior to November 14th, we used an RDD sample to get up-front commitment to fill out a mail questionnaire. We called 7008 numbers, of which 1831 were ineligible by virtue of being business or government offices, fax/modem lines, disconnected, beepers, etc. The remaining 5177 eligible numbers yielded 993 respondents who agreed to participate - or a yield of 19.2%.

On November 14th, the questionnaire was then mailed out to the 993 who agreed to fill it out. The package mailed was the same as that used for the control cell in Test I. As was the case with Test I, the fieldwork was concluded on December 19th. Of the 993 mailed, we received completed questionnaires from 659 - or a 66.4% return rate. The demographic profile of the respondents was similar to that for the other test cells. While one can also conclude the net response rate was 12.7%, we believe that for certain populations this method may have applicability, in that it provides a way to undertake at reasonable cost, research that would otherwise be cost prohibitive.

SUMMARY OF TESTS
Use of advance postcards and identification of a contact with whom survey validity can be confirmed increase response to mail surveys. While a telephone pre-recruit generates relatively low response rates in the initial phase, it yields high rates of cooperation in the second phase, and provides a relatively cost-effective way to survey low incidence populations or to conduct surveys where the cost/survey administered is unusually high. All other factors being equal (e.g., good lists, interesting subject matter, educated questionnaire design, etc.), we find the following:

APPROXIMATE RATE EXPECTED - ONE MAILING
5 Years Ago vs. Present

Additionally, we find:

• Two new single $1 bills are more effective than a $2 bill.
• Response rates will generally be 5-10% higher if the sponsor is identified rather than only providing the name of the research house (except with lapsed customers).
• An additional donation to a charity of choice can sometimes raise response rate 1-3%.
• Justified proportional typesetting is quite important as is printing on 100% rag white or off-white paper.
• A perceived importance to the sponsor (The Wall Street Journal, BusinessWeek, Glaxo SmithKline, etc.) enhances response rate and often negates the necessity of an advance postcard.
• Respondent interest in subject matter will increase response rates by 5-10% or more.

For further information, please contact:

Paul DiLillo (paul@nybeta.com)

John Martino (john@nybeta.com)