June 8, 2016
We are conducting The Client Trust Index™ to measure a company’s performance with its customers. They want to know what the customers really think and what specific behaviours they can change or improve to grow their business and deepen their relationships. We are surveying a random sample of their customers. The total sample size is 19,000.
My client asked, "how many customers do we need to participate in order to confidently make accurate inferences."
My response was, "384 is the magic number".
When conducting a survey, it is best to receive a random sample size of 384 respondents (n=384). When we receive valid responses from 384 people, we are able to achieve a 95% confidence level. This is a standard confidence level for social science research. The margin for error would be 5 %.
When determining a course of action, we want to be as confident as possible in the quality of our inferences. With a sample size of 384, we can describe, with confidence, a population in the hundreds of thousands, or even millions. Therefore, as we shared with my client, with a sample size of 384, we can confidently make inferences of the customers and the Client Trust Index™ level for their entire customer population of just under 90,000 customers. Many gallop polls and international surveys cite a confidence rating of 95% with a 5% error. If you look at the numbers closely, most of the time their sample size is 384.
A high confidence level will give you a solid benchmark for developing customer relevant strategies and plans. Here's sample size calculator, (http://www.raosoft.com/samplesize.html) have a look at it to see the sample size your company would need to achieve 95% confidence level.
If you are interested in learning more about how industry leading companies are using The Client Trust Index™ to gain insights to grow and adapt their business to better serve their clients or how to use data and analytics to make strategic data - driven decisions, take a look at The Client Trust Index ™ here.