11 replies to this topic

[Simon]

There is evidence to suggest that satisfied customers buy more often, they buy more individual product and more of your range, they will pay more (9% less price sensitive), they will recommend you more (positively) and they will stay with you longer. If the evidence is correct improving customer satisfaction should be a key goal for the business, it's also a requirement of ISO 9001:2000.

So how does your organisation measure customer satisfaction? What methods do you use? How often do you do it? Whose job is it? What is your target? Are you meeting / exceeding it?

Please tell us.

Regards,
Simon

[Simon]

I have a question. If I send a survey questionairre to 300 customers and I get 50 returned (return rate of 16.66%), how confident can I be the results are a true reflection of my customers perceptions?

I honestly want to know - can any statistics whizzos help me out please?

Regards,
Simon

[Simon]

I have a question.  If I send a survey questionairre to 300 customers and I get 50 returned (return rate of 16.66%), how confident can I be the results are a true reflection of my customers perceptions?

I honestly want to know - can any statistics whizzos help me out please?

Regards,
Simon

<{POST_SNAPBACK}>

I really would like to know.

[SAM]

We measure our customer satisfaction by using customer service levels (number of failed deliveries, problems with product, etc) and Complaints analysis.

Then every now and again we send out customer questionnaires. I have found that the best way of getting responses is by either sending the questionnaire when returning their Supplier questionnaire, sending them a simple form by email, or ringing them up and asking the questions over the phone.




Is there any reason why the banner at the top contains a link for "Manchester Psycotherapy....Professional therapy in Stockport"?

Is there something you want to tell us Simon?

Nadine

[Simon]

Hurrah! And hello Nadine, if this had gone on any longer I'd have needed a blooming shrink.

We measure our customer satisfaction by using customer service levels (number of failed deliveries, problems with product, etc) and Complaints analysis.

It's important to measure failed deliveries, problems with product and complaints but strictly speaking they're measures of your performance and not measures of how happy or unhappy your customers are with your performance. An absence of customer complaints is not a good measure of customer satisfaction. In most cases, dissatisfied customers don't complain - they just don't come back!

Every now and again we send out customer questionnaires. I have found that the best way of getting responses is by either sending the questionnaire when returning their Supplier questionnaire, sending them a simple form by email, or ringing them up and asking the questions over the phone.

I hope you get a better response to your surveys than I do. What do you with the results of your customer surveys? Are they discussed and action plans developed or are they used to wedge the canteen door open?

In answer to my question I found some information on the web and also on my web site:

Customer Satisfaction FAQ's

Statistical Confidence in a Survey: How Many is Enough?

Sampling Overview

Sample size calculator

Survey Random Sample Calculator

I used the above calculator and for a survey population of 300 I need:

176 responses to be 90% confident in results (4% error rate)
200 responses to be 95% confident in results (4% error rate)
233 responses to be 95% confident in results (4% error rate)

With my measly 50 responses the results are unreliable, but still give a good indication, well better than doing nothing.

Regards,
Simon

[Wallace Tait]

I have a question.  If I send a survey questionnaire to 300 customers and I get 50 returned (return rate of 16.66%), how confident can I be the results are a true reflection of my customers perceptions?Simon

The 300 customers is your population, 1/6th of your population made the effort to answer your questionnaire.
You can therefore be confident that the 50 customers who answered your survey are a true representation of a population.
Change your population: your new number is the 1/6th = 50, you can now make an estimation based on the 50 customers as to what would be the average answers of the 300 overall.
I would use the three most effective tools for measuring your survey, Run chart (Measure the response to each question), Pareto (Categorize according to importance) and Ishikawa (Correlate the root cause associated with the resulting answers).
If your survey is for CI purposes, I would say this is a balanced approach.
Wallace

[Simon]

The 300 customers is your population, 1/6th of your population made the effort to answer your questionnaire.  You can therefore be confident that the 50 customers who answered your survey are a true representation of a population.  Change your population: your new number is the 1/6th = 50, you can now make an estimation based on the 50 customers as to what would be the average answers of the 300 overall.

Hi Wallace thanks for answering, however, according to the research I did ‘statistically speaking' 50 out of 300 is not enough to be very confident in the results. Mind you I'm not a stats expert and I could be wrong.

Regards,
Simon

[Wallace Tait]

Hi Wallace thanks for answering, however, according to the research I did ‘statistically speaking' 50 out of 300 is not enough to be very confident in the results.  Mind you I'm not a stats expert and I could be wrong.

Regards,
Simon

<{POST_SNAPBACK}>

IMO, FWIW,
50 being your only available measure, would allow you to predict "statistically speaking" a reasonable measure based on your original population.
Wallace.

[Simon]

IMO, FWIW,
50 being your only available measure, would allow you to predict "statistically speaking" a reasonable measure based on your original population.
Wallace.

OK and your opinion is worth a great Deal Wallace.

To be honest I'm happy with 50 out of 300 it will give a feel for customers perceptions. Sometimes I think we get hung up on the accuracy of analysis to the 'nth degree' and it often blinds us to the most important thing, which is doing something with the results.

Thanks for your input.

Regards,
Simon

[Wallace Tait]

I think we get hung up on the accuracy of analysis to the 'nth degree' and it often blinds us to the most important thing, which is doing something with the results.

Yeah you're right.
It has been said that "statistics can be 99% fact and 1% truth".
I may agree more often than not.
Wallace.

[Simon]

Another question on customer satisfaction. Who monitors CS in your company, is it marketing, quality, a bit of both, or someone completely different?

Regards,
Simon

[Wallace Tait]

Another question on customer satisfaction.  Who monitors CS in your company, is it marketing, quality, a bit of both, or someone completely different?

Regards,
Simon

<{POST_SNAPBACK}>

Marketing and quality metrics are so related to each other that they are treated as a single unit, within the auto industry for the most part.
Warranty issues are a main measure for CS.
Wallace.