Sample Size Calculator for Contact Centers

How many phone calls do you need to evaluate each month for a statistically significant sample?

Are you tempted to just pick a number out of thin air?

Is it the same number for each of your reps?

Let’s get real for a second.

If you don’t evaluate enough calls for each rep, then you don’t actually know how your employees are representing your business.

Use our calculator below to figure out exactly how many interactions you need to monitor for each agent, then keep reading to understand why and how.

# of calls to evaluate per agent each month:

Sample 100% of your agents’ interactions with Voxjar's AI Evaluator.

How are we calculating the sample size?

If you want to calculate your sample size by hand the formula is:
p = Population (call volume/# agents)                      e = Margin of error (we default to .05)                 z = Z-score
A few things that will impact your sample size:
• The smaller your margin of error the larger your sample size will need to be for the same population.
• The higher your confidence level needs to be, the larger your sample size will need to be, too.

How do I determine Population Size in my Contact Center?

In the case of contact center Quality Assurance, each agent has their own population. It's the total number of calls each agent makes during the period that you're evaluating. In the calculator above we do a rough estimate for you by dividing the call volume you provide by the number of agents.

It's important to split your total call volume by individual agents because you're evaluating them on an individual level. Each rep is going to interact with customers in their own unique way.

Lumping all of your agents together is tempting, especially since it usually decreases your sample size, but you'll just be tricking yourself into thinking that you have a good overall view, when in reality, you'll be over-sampling some of your agents and under-sampling others.

What is Sample Size?

Your sample size is the number of phone calls your quality assurance team will need to monitor to get a statistically significant view of your agents' interactions with your customers.

Ideally, you'll determine a sample size for each individual agent since volume (population) will vary from rep to rep.

Sampling is necessary when your call center uses traditionally quality assurance because the cost to manually evaluate every call would be outrageous. Sampling scientifically, randomly and with the correct sample size,  is how you achieve statistical significance when evaluating and improving your reps' performance without an automation software to help.

To eliminate bias in your sampling it's important to do so randomly. Use either a quality assurance software with built in random sampling or create your own randomizer in excel

Which Confidence Level should I Use?

The higher the confidence level the larger your sample size will be. If you choose 95% then your sample size will adjust so that you can by 95% certain that the population is represented in your sample.

It's not the most intuitive measurement. You can read more about it here: Confidence Level

My Sample Size is HUGE!?!

That's not really a question, but I get what you're saying.

Most call centers are massively under-sampling their agents' calls.

Not only are they missing out on opportunities to improve their agents' performance and company revenue, but they are also opening up their businesses to major liability.

Imagine evaluating 8-10 calls per month for an agent that takes thousands and thinking everything is on the up and up for three or four months.

Then, in month five, you come across this agent lying to a customer or mishandling credit card information.

Now you're in a pickle.

Was it a one time thing? How long has this been going on?

Do you go back and listen to hundreds of their calls to see if this is a long term problem?

If it has been going on for months, how much damage has it caused?

This happens all the time. Most of the time the rep gets fired and the issue gets swept under the rug. Because what else are you going to do?

Well, it doesn't have to be like that anymore.