Small Business Survey Sampling

To perform customer satisfaction research, you need to make sure you have an adequate sample size. For most larger businesses, this is easy, because you simply need to randomly poll just a small fraction of your overall customer base on order to receive an adequate sample.

But for smaller businesses this can get somewhat tricky. Most survey research by large companies tries to get a sample size of about 1000 customers (1058 is the most common number) because this provides a 3% margin of error when the researcher is looking for 95% confidence.

But larger companies also have hundreds of thousands of customers to use in the sample. Smaller companies only have about 1000 to 2000 customers at most. While customer satisfaction is still important, obtaining a sample of 1058 respondents is far more difficult.

Indeed, by that point you may as well sample your entire population, and most likely that is not possible because response rates are going to be difficult to obtain. So what does this mean for survey research? Obtaining a sample at a 3% confidence interval still requires a fairly large sample – anywhere from 500 to 700 respondents depending on the number of customers.

For companies with only 2000 customers, that’s would require you to reach essentially every customer and hope that you receive a large enough sample. However, if you drop that number to a 10% confidence interval you only need roughly 90 respondents. Small businesses may need to sacrifice the value of their data in order to make their research easier.

If you’re researching customer satisfaction, one might argue, however, that sample sizes do not need to be as large as they do with larger companies. It may depend on what you research.

For example, if you are trying to improve satisfaction with your current customer base (which is not very large), and a large percentage of them respond to your survey with an idea that will improve their satisfaction, it may not be as important to get a larger sample, because now you know exactly how to target the satisfaction of those customers. Often when your customer base is small, the reality is that you simply need to “reach as many customers as you can” without worrying about specific numbers. While this is not necessarily a good idea mathematically, the reality is that the limitations of your population is going to prevent you from reaching an adequate sample anyway, but the data you receive from that sample should have more meaning (not necessarily statistically, but practically) to your business.

Only when your customer base has increased to more than 2,000 customers should you start to worry about reaching the sample size of larger companies.