Market research has long dominated as a decision-making tool for broad-based marketing. Indeed, they were the only way to collect data on the behaviours, expectations and opinions of consumers. The mathematical techniques of data analysis and modelling in vogue since the advent of “Big Data” have in fact been used for a very long time on study data, to offer powerful tools to the decision maker: barometric monitoring of satisfaction, detection of typologies of customers, tests of concepts, products, packaging and advertising, etc., with great successes!
Nevertheless, these techniques have always had their disadvantages, among which are the difficulty in obtaining representative samples, the multiple biases inherent in any “declared” data, the lack of depth in the data (or sometimes too high costs) , and the general impression of a given may be too subjective.
The emergence of solutions capable of collecting and crossing multi-source data
Faced with this, the emergence of robust customer database (e.g., CRM) systems with increasing data quality and the ability to link previously separated data sources within a Big Data ecosystem bring a new dimension to consumer analysis, a global vision in which all consumers are registered with their purchases, their browsing behaviors, their profiles on social networks, and for some businesses, their contract renewal dates. This data avoids the pitfalls of the studies since it has a much greater depth, if not complete, that it is a truly objective (undeclared) measure of behaviour, and that it is rich in diversity that will only with the opening of data via Open Data.
The only way to collect the “why” of an action is to ask the question
However, the promises made by a relegation of studies too heavy, too old, too slow, to forgetfulness undoubtedly went too fast in their conclusions. If the superiority of the CRM datum in the broad sense of observing all that is observable in the entire world is undeniable in relation to the studies, it will always lack this access to the human, its will, to the “why” of which he undertakes. Now, as long as technologies capable of reading our intentions directly in our brains remain in a fantasy state, the only way to collect the “why” of an action is still to ask the question.
We will be able to achieve a high performance attrition score based on CRM data, without understanding what drives customers to leave, let alone analyzing the different causes of departure: dissatisfaction, best competitive offer, or customer satisfied but simply does not need the service? It goes without saying that, in addition to an effective targeting of “churners”, the reason for attrition must also be understood in order for any retention attempt to succeed because it must take account of expectations ( or lack of expectation) of the consumer.
CRM data and study data: each brings a different light on the customer
Thus, there is no one method superior to another: there are different data that have their dedicated means of collection, and that each brings a different light on the customer. If one were to attempt to categorize the “ideal” uses of these two types of data, one could describe the CRM data as a factual and complete observation of the customers which allows to detect and to anticipate its future behaviour in order to take actions highly operational and potentially highly targeted marketing. The study data, on the other hand, makes it possible to understand what motivates the consumer in his actions and what they think, although not exhaustively on all the customers, but sufficiently to define the content of the marketing actions that the consumer, we will undertake thereafter. To summarize,
If it is understood that the uses and the purpose of these two types of data are different, does this mean that they must be separated from each other? Are there not opportunities to gather these different data corpuses?
Data Expanding, the art of knowing what each of our customers thinks
The collection of study data is inherently limited in depth. It is impossible to probe the whole of a client base (with questionnaires sometimes very long). This is where the so-called Data Expanding techniques come in. It is based on a sub-population of customers polled and which can be found in the client database (via the email sending the questionnaire for example), to select the variables from the questionnaire, modelling them from the CRM data common to the individuals surveyed, and then applying the model to the entire database. In this way, it is possible to know what each of our customers thinks.