Expanding Your Customer Base

Why Sample Size Is Important For A Consumer Insights Survey

9 November 2022

Jorge Medina

Every consumer insights service provider will hammer the importance of robust sample sizes. However, this leads clients to wonder what the ideal size for a sample might be. There isn't a one-size-fits-all solution, but there is a way to arrive at an answer that addresses key issues. Let's look at why sample size is important and how to determine the right size for your survey.

Population Size

The biggest factor that dictates the necessary sample size is the size of the target population. Smaller populations require bigger samples to be representative. If you're surveying a population of only 1,000 people, you'll need to get answers from at least 300 of them to have a dependable read.

As the population size goes up, the acceptable percentage goes down. 10 percent is a commonly accepted sample size for a population of 10,000. This is why you'll see political pollsters be happy with something like 1,300 respondents in a state with a population of 10 million people.

Why is this the case? Large populations tend to be more homogenous, and randomly sampling a small number of their members is unlikely to rope in lots of outliers.

Smoothing Noise

The larger the sample is, the less noisy the data will be. If a consumer insights survey service firm sampled two random groups of 10 people, for example, the differences of opinion between the two groups could be radical. A change in one person's opinion, after all, could represent a 20 percent swing in either direction. That is way too noisy. However, 100 people will create a much less noisy result.

Notably, there is always a limit to the benefit of going up further in sample size. The noise reduction going from 100 to 1,000 respondents will be noticeable. Going from 1,000 to 10,000 will likely be a waste of time and money.

Diversity and Subsamples

One reason to go with a much larger sample is that you're concerned about representative subsamples. If you want to have a representative subsample for a minority group that represents 10 percent of the population, you may need to increase the sample size to guarantee representative results.

Consumer Modeling

Once you have the data, you need to consider how it might match your model of the consuming public. If you know from many other surveys that 75 percent of your product's buyers identify as female, for example, you may need a larger sample to pull in more women if you expect something closer to a 50-50 split among respondents.  

Share

Search