There are many myths about NPS that, as the reactions of my blog readers show, many take for truth. I will try to dispel some misconceptions.
There are many different ways to get customer feedback. Various indexes, techniques, etc. are used. Many people think that the NPS index is the best method and prefer to work only with it. But this approach is wrong. NPS is not a universal tool for measuring customer loyalty, it is just one of the indices.
Client point of view
And the oldest, but not at all obsolete index is CSI. The acronym stands for customer satisfaction index. In Russian translation, TZK is often used – the client’s point of view. The CSI is familiar to many. Often at gas stations, in banks, in stores at the checkout counters, you can see a panel, as in the picture below. The main question for determining the CSI is “How satisfied are you?” There are several answers to it, ranging from “dissatisfied” to “satisfied”.
Although this method is not very informative, it is easy to use. CSI surveys are essentially a convenient format for finding problems. Such surveys are also good because they can be used to evaluate goods and services, the work of a service or personnel, the work of a website or a call center. You can evaluate any aspect of your business.
For example, at the end of this post, I can ask readers if they liked this post and my blog in general. Separately, I can ask them about the design of the blog, its theme and structure. And in order to get feedback from readers, you don’t need to use the NPS index. As I wrote before, NPS relates to the overall performance of the company.
Here are some examples of where the CSI can be applied.
- Product. Find out how your product meets the expectations of consumers, whether it satisfies the quality, whether it is pleasant to use, color, sound, feel, etc. In a word, everything that is most likely important for you to know about your product.
- Services / service. By analogy with the previous point, you can ask about any aspect.
- Staff work. Ask how well a manager, call center, or tech support has done.
- The interface or content of the site , the operation of the mobile application or the order form on the site. As well as the relevance and attractiveness of sending by mail or SMS.
That is, any aspect of work can be assessed using a simple CSI survey. For example, you can find out the attitude of respondents to such criteria as the speed of response to a request, the color of the uniform, the taste of pies in your cafe, or even coffee from a vending machine.
How many questions at a time
Of course, do not forget about common sense. And although you can ask the respondent to rate several aspects at once in one survey, it is still better to limit yourself to 4-5 points. Judging by the practice of my work, in cases where there are more questions, the curve of the number of answers goes down sharply. And this is a very important nuance. If the number of answers is less than 30-40%, this means that the survey is not conveniently constructed, the questions are asked at the wrong time, or you actually have a huge number of very dissatisfied customers who are sorry to put even one. Another option: perhaps the question is asked in such a way that the client simply does not understand what they want from him.
It is optimal to ask three questions at a time. Thus, you will not irritate the client and will receive information for analysis. It is extremely important to feel the measure. If the user is asked to participate in a survey after each of his actions (online or by phone), he will very quickly begin to avoid them. This is understandable, because sharing his opinion is not his job or duty, but a favor done of his own free will.
How to measure CSI
There is no universal scale. However, the calculation of this index is still easier than in the case of NPS. It may be the usual five-point system, or it may be ten or even three points.
When choosing a measurement system, consider the following:
- The more divisions, the less precision. Not everyone is able to understand what is the difference in “satisfied by 6, 7 or 8 points.” And even if the respondent gives exactly 7 points, then what to do with this result? I always urge you to see real people behind the numbers and clearly understand how you can use the survey results. How important are these shades of satisfaction? If a score of 6 or 7 is important, of course a 10-point scale should be used. However, it is always best to use a number of points where the numbers can be replaced with a sentence that will help the respondent understand the essence of the question. In this case, it is easier to analyze the obtained data. Here is a more or less standard scale for five points.
1 – completely dissatisfied
2 – rather dissatisfied
3 – satisfied with something, not with something
4 – rather satisfied
5 – completely satisfied
With three points, everything becomes even easier:
1 – dissatisfied
2 – ok
3 – excellent
It is easier for respondents to answer, of course, using a three-point system. Therefore, the choice between three- and five-point systems is usually made by the researcher himself, based on how he is going to analyze the results. For example, if you need to display the average index and track its changes, then a three-point assessment is enough. If you need to understand the client’s attitude more clearly and in detail, then you need to take a more detailed rating scale.
I recommend to all my clients to give up such a familiar and convenient “middle”. In our culture, it is unacceptable to complain, so the respondent can simply put a three and say “okay”. But this is not enough for those who need a real result and work with an index, and not just an average index. Then it is better to remove the three and use the 4-point system. An even number (you can use a 6-point one) will more accurately and clearly show the results. This is how you can induce customers to choose, based on the logic that they are more likely satisfied or rather dissatisfied, without a convenient “C”.
Where and how to use CSI
The biggest advantage of CSI surveys is their simplicity. With their help, you can appreciate any stage of interaction with the client. This allows you to quickly find weaknesses and also quickly work them out.
If the survey gives the same scores over a period of time, you can end it. This suggests that the criterion under study is debugged and working well, then another aspect can be assessed.
For example, if the assessment of the convenience of a mobile application within 3-4 months gives the same high results, then the application is really user-friendly. As long as no changes are made to the application, you should not ask about its usability. In general, changes are always a reason to ask customers “how do you like it?”
In some cases, for example, you can find out the attitude towards change:
- Your service has moved
- The sales manager of the car dealership or the master inspector in the service has been replaced
- The contractor of the call center changed
- The site or application has changed design
- The conditions of ordering or delivery in the online store have changed
- …
I deliberately cite both particular questions and generalized ones, so that it is clear that any changes are a reason for a new assessment.
I also wrote about other indexes, for example about NPS:
The NPS Consumer Loyalty Index – what is it?
So why do you need CSI? For operational management, to identify problem areas and respond to them. Although this is called the satisfaction index, in fact, its main task is to find weaknesses, to identify what customers are unhappy with, and at what stage of interaction is the weak link. And of course, to assess changes or some kind of innovation, yes, even a new product in the end. Yes, CSI will not say anything about supporters and critics, about customer loyalty (because satisfied does not mean loyal). But CSI will provide much-needed information about a specific problem area. That is why we should not forget about this index, and that is why in the world research of the level of satisfaction, this index is ahead of the NPS in popularity and frequency.
