Depending on the informational objectives, available budget, and expected timelines, traditional market research can be conducted using a variety of different techniques. Typically, the best and most comprehensive results are achieved using a two-phase approach, based on two levels of investigation to be developed sequentially:
- Qualitative investigation
- Quantitative investigation
The qualitative investigation aims to identify the most relevant elements (attitudes, desires, preferences, value systems), while the quantitative investigation aims to measure the weight of each of these elements. This type of approach can yield good results but has some basic structural limitations, primarily related to time and implementation costs. We discussed these limitations in this article.
Regardless of whether the two-phase method is used or not, some critical points emerge at a methodological level and must always be taken into consideration. Whatever research method, it will always be based on two key elements:
- A sample
- Some interviews
These basic elements make research economically sustainable and physically manageable within acceptable timeframes, but they also introduce critical elements affecting the significance of results and the time and costs of implementation.
The use of a representative sample of the target population makes it possible to conduct interview-based research even in the case of large target populations. However, there are some points of concern:
- The margin of error
- The sample size
- The impact of the sampling method.
The statistical error
The statistical error is the mathematical expression of the expected level of precision of the results measured on the sample compared to those measured on the entire target population. It practically measures the level of representativeness of the sample. The higher the level of statistical precision desired from the analysis, the larger the required sample size: to halve the error, the sample size must be quadrupled.
The sample size
In addition to being the inverse function of the statistical error, the required sample size is also the direct function of the complexity of the target population to be represented: the more complex and composed of sub-targets to be investigated, the more they will need to be represented in the sample, increasing the sample size.
The combined effect of accepted error and target complexity can have explosive effects on sample size, construction complexity, and difficulties in recruiting interviewees. Ultimately, this impacts time and costs.
The sampling method
The term “sampling method” refers to the selection of subjects to be interviewed. There are two methods:
- The probability method
- The non-probability method
The probability method involves the random selection of the predetermined number of subjects, according to the established structure of the representative sample of the target population. This method allows for the a priori calculation of the statistical error. However, for the selection to be truly random, it should be made from the entire target population, which raises two basic problems:
- Availability of contact data: in reality, the research institute never has access to the entire available population.
- Geographic distribution: random selection results in an equally random dispersion of the interviewed subjects across the territory, which often makes in-person interviews unmanageable.
The non-probability method involves the direct selection of interviewees by the researcher. This eliminates the problem of dispersion across the territory but removes the requirement of randomness, making the statistical error effectively incalculable.
Interviews can be conducted in person, over the phone, or via the internet. There are some types of distortion inherent in the method of surveying through interviews. In summary:
- The interviewer. Their presence can induce the interviewee to provide untruthful or partially altered responses that are pleasing to the interviewer. In addition, the interviewer’s questioning style and attitude can alter the interviewee’s responses.
- The social conventions. The interviewee tends to provide responses that give a “socially acceptable” image of themselves.
- The environment. Interviewees may, for example, not want to talk about certain topics when interviewed at home, or not want to spend too much time on the phone or the street. They may not pay enough attention to completing questionnaires on the web because they are distracted by other elements.
- The conduct of the group or individual interview. The specialist’s mode of conducting the dynamics of the group or in-depth individual interview has a significant influence on the course of the discussion and the elements that emerge.
- Group dynamics. During group sessions, it is necessary to pay close attention to obtaining a balanced discussion and the emergence of all points of view. This requires careful management of the dynamics between emerging leaders and less willing and more closed-off participants.
Overcoming Limits: The Internet
The evolution of the internet, along with the explosion of online publishing, blogs, and social media, has created a digital square where people share an enormous and invaluable amount of information about their experiences, aspirations, evaluations, and ideas.
It is an immense informational heritage: to get an idea, just consider that every day, more than 4 billion people connect to the internet, generating about 7,500 terabytes of new data.
Welcome to the jungle
Reading and interpreting discussions, posts, articles, tweets, and any other form of expression allows us to access the information we are looking for, bypassing the limitations of the sample and the distortions introduced by interviews. We no longer question some people to represent others, nor do we subject them to any conditioning. We simply observe and listen to all of them.
It’s a bit like studying the behaviour of a tiger: by observing it at the zoo, we can get a rough idea, but it’s only in the jungle that we see it in action.
RTBH.ai does exactly what we need: it reads, interprets, classifies, and synthesizes everything expressed online related to a brand, product, or market. It provides timely and actionable information.
- It does not analyze a sample, but the entire universe
- It does not introduce any conditioning toward the target
Its core is a proprietary artificial intelligence algorithm developed in collaboration with some of the world’s top universities. Its uniqueness lies in combining in a single tool:
- Artificial intelligence
- Machine learning
- Cognitive sciences
- Data science
- Social network analysis
RTBH.ai scientifically and accurately measures the health status of any brand and its competitors, synthesizing it thanks to the proprietary “Brand Health” indi