Customer Satisfaction Survey Sample Size

Hi!


In our case it's not possible to administer questionnaires for the customer satisfaction survey to all the customers.
How do I choose the sample size on the administer the questionaires?

Thanks
AV

António Vieira said:

Hi!


In our case it's not possible to administer questionnaires for the customer satisfaction survey to all the customers.
How do I choose the sample size on the administer the questionaires?

Thanks
AV

Click to expand...

Actually it is a two-fold question I would think. First is how many and second is who are those customers that should get it. How much precision do you need in your estimates?

An interesting phenomenon that you can usually observed from things like this is the pareto principle. Somehow, you'll find that about 20% of the customers gave you 80% of your business (in terms of business volume). If you drill down further, you'll noticed that about 5% (20% of 20%) gave you about 60% of your business volume.

I'll accept if those replies account for more than 60% of my business volume rather than focusing on the number of replies.

António Vieira said:

Hi!


In our case it's not possible to administer questionnaires for the customer satisfaction survey to all the customers.
How do I choose the sample size on the administer the questionaires?

Thanks
AV

Click to expand...

What do you want to know?

Surveys have a notoriously low return rate, and too often are skewed because unhappy people tend to say more than happy ones.

This is why I usually recommend the person closest to the customer be trained to probe for satisfaction (as casually as possible) and be empowered to remedy imperfections on the spot, perhaps even going into discount territory if practical.

harry said:

An interesting phenomenon that you can usually observed from things like this is the pareto principle. Somehow, you'll find that about 20% of the customers gave you 80% of your business (in terms of business volume). If you drill down further, you'll noticed that about 5% (20% of 20%) gave you about 60% of your business volume.

I'll accept if those replies account for more than 60% of my business volume rather than focusing on the number of replies.

Click to expand...

I also support this concept. Ask yourself which few customers make up the Lion's share of your business, then go from there.

how about the other ones.
I would say the organization has to focus on long term customers instead of the volume.

amanbhai said:

how about the other ones.
I would say the organization has to focus on long term customers instead of the volume.

Click to expand...

This may appear to be true on the surface but in reality, every company has and should have various
categories of customers - cash cows & dogs (long term customers as you mentioned) as well as stars &

question marks (new & potential). Refer to the BCG matrix also known as Growth-share matrix for an idea as well explanation.

Our problem is not the low return rate of questionnaires (in our business we can have 100% of customers replying to the questionnaires).
The same way measuring devices must be calibrated in order to assure reliable measurements, customer’s questionnaires surveys should be also reliable. In other words we need to know the internal consistency of each dimension stated by a group of questions in the questionnaire.
Following this idea we need to make a test on the questionnaire reliability before we submit the questionnaires to all the customers. That’s why we need to know how many customers we need to use in this pre-test of reliability.
If we have for instance 1000 customers in 20 different places (50 in each place), should we use for example 10/5 (20%/10%) customers in each place for this test?