Predictions in market research play an important role for all companies. It is clear that we can only be certain, and to a certain extent, about consumers' past behavior. We can inquire about their past actions with a high level of confidence. To the question " What did you buy yesterday? " we assume that the answer, if there is no intention of the interviewee, will be 100% certain. If we go further back in time from the purchase act under study, for example:
What did you buy in your last 10 purchases? Then it is to be expected that the probability of certainty in the answer is no longer 100%. And yet, we know from studies that compare actual behavior (purchase panel) versus the perception of what was done (survey), that people confuse their past actions and do not always agree on everything, obtaining differences that range between 15 and 20%.
However, even assuming these limitations, prediction and forecasting models are a fundamental part of market research and are an integral part of the innovation processes or funnels of large companies.
I recommend you read: The importance of context in a product test .
Market research to measure
How can we know in advance whether a new idea will achieve the belize phone number expected sales volume? There are different parameters or algorithms usually used by market research teams and in predictive innovation models.
In some cases, expectations are assessed simply from results on certain parameters taken as key performance indicators. For example, traditional scales of purchase intention . Based on the percentage of people who agreed “that they would buy” the product, we can infer a certain potential for the launch.
Across categories, targets, and cultures, these reference values change depending on whether the top or top two box is used . The way respondents use these scales varies greatly depending on the culture and universe in question.
I invite you to also read: Measuring purchase intent with a focus funnel .
It is very difficult to standardize criteria. In fact, in certain cases, in light of the results, the “strong desire” of Brand Managers to continue moving forward with the project leads to the relaxation of certain previously agreed success criteria.
In companies with a long history of innovation that conduct market research, these parameters tend to be more solid and are based on extensive data of experiences, forming “norms” for reference. Instead of being an almost arbitrary number that is valid in itself, a historical average of similar experiments is used.
The validity of this reference criterion requires repetitions of experiments under similar conditions, same or adjacent category, same target universe, identical question structure, etc. Building a normative base is a systematic task that requires investment.
Clear and specific protocols are needed: evaluating with a brand is not the same as evaluating without a brand, with a price or without a price, a complete or incomplete mix. In my career as a researcher, I have seen several companies make this effort to build standards for years before finally being able to see the fruits of their work.
You may be interested in reading about market and customer needs research .
But even then, our certainty will not be 100%. The conditions of an experiment are rarely exactly repeatable. We can control the demographic aspects of the universe, but not its attitude profiles, its relationship with the category and the brand.
- Board index
- All times are UTC
- Delete cookies
- Contact us