When we manage a service, develop a Czech Republic WhatsApp Number List and seek to improve it, Lean 6 sigma teaches us to act according to what the voice of the customer tells us. For that, it is necessary to listen to it and to rationalize its requests, that is to say to translate quantitatively its needs (physical sizes, count, etc.). However, it is not possible to measure everything, to control everything, and after all – the customer being king – the only real indicator that ends up being really valid tends to be his satisfaction.Personally, that scares me a little because satisfaction cannot be measured with a thermometer, a voltmeter, or a sound level meter
but by survey. This is a measuring instrument like any other except that to compensate for the lack of a common reference frame of “contentment” between the respondents, we rely on the quantity of responses (and some statistical methods). to produce objectivity. So, at the risk of sounding cynical, a customer’s voice isn’t really worth much; the voice of the customer (that is to say the sum of the customers) on the other hand… My aim in this article is to establish to what extent THE customer is able to judge for himself what is important for him in a service / product.
The voice of the customer
Throughout the article, I invite you to discover the methodological elements on which what I am putting forward is based. Reading these elements is not necessary to understand the key findings , I nevertheless remain interested in your contributions for this method! For the demonstration, we will use as a common thread, the survey carried out by ARCEP relating in particular to the satisfaction of decision-makers in SMEs / mid-cap companies with the Internet access provision service . In this survey, the satisfaction sub-criteria present to detail the overall satisfaction with the Internet access provision service are as follows:
The importance of the sub-criteria of the internet provision service as expressed by the respondents The importance of the sub-criteria of the internet supply service reconstituted by multiple linear regression and materialized by the coefficients (weight in the equation) in front of the satisfaction variables on each sub-criterion (cf. equation of the “Method” diagram) What customers are saying Let us observe the average importance given by respondents to each satisfaction sub-criterion; for each sub-criterion, the respondent could score from 0 to 10 his personal estimate of the importance of said criterion. The following results are obtained:
I did not understand you
This is a typical case: when you ask clients what is important to them, the result is a form of white noise where each criterion is important (6-8) or very important (8-10). The purpose of a survey being to identify the most relevant actions to be taken to improve overall satisfaction, these results do not help us so much … By testing several models to predict the overall satisfaction from the satisfaction on the sub-criteria (linear regression, Random forest, etc.), I obtain the best performing model to predict the overall satisfaction. It is the following (multiple linear regression):
The reconstitution of the importance of the criteria by calculation made it possible to avoid the “white noise” effect: it reveals criteria which are moderately important (4-6), not very important (2-4) and not important (0- 2) for the customer. The order of criteria classified by level of importance is no longer the same (see table below); eg: the upward flow criterion has changed from “very important” according to the respondents to “not important” according to the calculation . This is interpreted by the fact that in reality, there is a weak correlation between overall satisfaction and satisfaction on the “upward flow” criterion.