Research 101: Estimating Telephone Survey Research Costs

By Ted Apostol and Rebecca Irvine

Determining the cost of a proposed research project can be difficult — and painful — especially if the project calls for new parameters and strategies. However, knowing the basics of estimating marketing research costs can help set budgeting costs and plan RFP needs for future projects to be completed.

Research costs are almost entirely in terms of man-hours. Research companies estimate the hours needed for the various tasks, multiply the hours by billable rates which include overhead costs, add materials, and the sum is the project cost.

There are basically three variables in the mix of research costs: the number of interviews, the length of the interview, and the report. These variables are attached to the following research tasks to determine cost.

• Sample size: Generally speaking, the larger the sample the more expensive a research project becomes. However, sample size can be relatively small due to statistical laws. In any given population, a sample of 400 is frequently recommended because it keeps costs reasonable while providing acceptable precision levels — plus or minus 5% at 95% confidence. While larger sample sizes increase precision, the cost may not be worth it. A sample of 1000 may cost twice as much but only reduce statistical error from 5% to 3%.

• "Incidence" of qualified respondents in the population: The more difficult it is to locate qualified respondents to take a survey, the more screening time is spent in the phone center. A low incidence target population can dramatically increase study cost.

• Length of the survey: Longer surveys, of course, take more interviewing hours, but they also increase the time for coding, data processing, supervision, and analysis. Surveys longer than 10 minutes make it more difficult to get respondent cooperation affecting both cost and representativeness of the sample.

• "Open ends": Questions that are open ended (that do not have specified responses to choose from) take more interviewer time and add time to code responses into categories. The typed feedback will increase the length of the survey and require a person to read and "code" all responses in a meaningful way. Open ended questions in which many of the answers can be anticipated can be pre-coded which keeps costs down, however open ended questions such as "How would you describe WestGroup to a colleague?" are more labor intensive.

• Report: The cost to analyze a study is mainly a function of the interview length. The more questions there are to analyze, the more analysis time is spent. The expected detail or number of visuals also affect the cost. Full-service analysis usually takes the analyst two to three hours for each minute of the interview length plus two or more hours for meetings and phone calls with you. (A 10-minute interview will take an analyst 20 to 30 hours to create a report.) You can save money if you trim the interview length or if you want just basic tables and graphs.

Full Service Telephone Study Cost
Typical Study of 400 General Public

The graph above shows the cost for a typical telephone study of 400 respondents at various interview lengths in minutes. You can get a pretty good ballpark cost for a telephone study using this graph. Remember this is for an average study with little or no screening of respondents. Population incidence and special needs could affect the cost significantly.

MAY 2000 Answers:
Home | Previous
Past Issues