I wondered what an acceptable number of surveys is? I work for a major bank and we are told they only need to have a rolling number of 7 calls to make the NPS score “valid”. Thsi seems an extremeley small cross section of the clients you meet. Also can you advise is it normal that cleints are called prior to a transcation being completed i.e a home loan this is only half done?

Cheers

Greg

A sample size approach is where you pick a representative sample of your population and conduct a survey. I presume your question is more about what is a census approach. In a census approach you survey all customers. Of course this depends on the application, industry segment and type of NPS survey. Let me explain.

There are two types of NPS surveys. The first is called a Relationship Survey and measures organisation wide performance. Typically this is conducted periodically. For our clients we suggest 6 monthly intervals This method typically uses a sample size approach similar to a Brand tracking survey. A sample size approach is particularly valid in a B2C situation where there may be thousands of customers. However in a B2B application where there may be only a few hundred customers it may be valid to go to all your clients even for a Relationship survey.

The other type of NPS survey is an Operational NPS survey which is what I was referring to . This is conducted after a ‘KEY’ interaction only and is conducted on an ongoing basis as opposed to periodically. In this case a census approach is better suited especially if it is after a key interaction. Of course you need to apply judgement. If you contact all customers after a key interaction you can ensure that you mazimise your NPS score by following up detractors and attempting to convert them. Plus it provides front line staff with regular robust feedback that is good and constructive enabling continuous improvement at a granular team level.

I have seen situations where a bank branch is given 3 completed surveys per week – 1 for Tellers, 1 for Customer Service and 1 for Lending. They are then told to workshop improvement strategies and obviously 3 surveys is not enough to enable business improvement. In this situation a Census approach with sensible contact management rules would make a lot more sense. This is why I suggested a census approach for operational surveys.

Hope this is clear. If you have any further questions let me know.

What do you mean by a census approach vs a sample based survey approach?

Thanks!