Define the target of your market research
Everything you need to know about targeting, representativeness and sample sizeIt is better to target too broadly than not enough
During the first stages of a project, it is important to keep an open mind. It is better to question a large target and be able to « zoom in » on a more specific core target when analysing the results, rather than directly question a target that is too small.
EXAMPLE
Project for an organic make-up line sold in parapharmacy
- core target group
- women 25-45 years old, users of organic parapharmacy make-up 2 to 3 times a week
- large target
- women 18-65 years old, make-up users at least once a week
If you only question your core target, you will have no way to assess the overall potential of your idea and your financial extrapolations will be less accurate.
EXAMPLE
Let's take up our project of an organic make-up line sold in parapharmacy. If during your study, 23% of the women surveyed say they want to buy your make-up line, how can you extrapolate the results according to the target audience ?
- core target group
-
women 25-40 years old, organic parapharmacy make-up 2 to 3 times a week
Two problems arise :
- The size of this population is not known and can only be estimated from more or less reliable documentary data.
- It also does not take into account the product's potential outside your core target audience (how many women under 25 or over 40 years old could you reach with your product? What would they represent in terms of turnover?)
- large target
-
women 18-65 years old, make-up users at least once a week
This broad target has several advantages :
- The penetration of make-up is known and comes from a reliable source: 64% in France (source : LSA-conso)
- The population of women aged 18-65 is also known via data from Insee : there are 19.7 million women aged 18-65 in France.
The potential of your market can therefore be calculated reliably :
19.7 million women 18-65 years old x 64% of make-up users x 23% interested in your project = 2.9 million women
In addition, if the results of your study are not very favourable, you will not be able to know if the concept was really not viable, or if it was only a targeting error.
EXAMPLE
If I only asked my target audience, how do I know if my organic make-up line is really not original and attractive enough ? Wouldn't it have been better for 40-65 year olds ? Could it have seduced non-organic makeup consumers ?
The importance of representativeness
Often, your target audience of potential consumers is thousands of people. It then becomes almost impossible to interview the entire population. This would be too long, too complex and too expensive. It will then be decided to interview only a part of this target via a sample considered representative.
It is only under this condition of representativeness that the results of the market research can be extrapolated to the whole of your target audience, and used to make financial and commercial forecasts.
Composing a representative sample of a target means ensuring that the fundamental characteristics of its target population are present in the sample in identical proportions.
EXAMPLE
19% of the French population is over 65 years old. If we wish to interview 1000 people representative of the French population, we will have to interview 190 people over 65 years of age in our sample.
These fundamental characteristics that will be controlled to construct the sample are called « quotas ». They generally concern demographic variables: gender, age, region, socio-professional category…
Interviewing consumers selected by online panels as proposed by poll&roll guarantees you a quality sample that is representative of your broad targets.
Which sample size to choose ?
The sample size is the number of people who actually answer your questionnaire.
The optimal sample size is a compromise between :
- the size of the target population. The ideal sample size changes little for populations of more than 10,000 individuals.
- the margin of error is an estimate of the extent to which the results of a survey can be used if the survey is repeated. In most cases, a margin of error of 5% is accepted. It is not recommended to choose a margin of error greater than 10%.
- the level of confidence expected. That is, the probability that the responses of the sample will reflect those of the population from which it comes. The most commonly used levels are 90% and 95%. It is not recommended to go below 90%.
EXAMPLE
For a sample with a margin of error of 2% and a confidence interval of 95%: if 76% of respondents say they are interested in your product during the study, you can be 95% sure (confidence interval) that between 74% and 78% (76% +/-2% margin of error) of your target will be interested in your concept.
To go further....
Good representativeness and optimal sample size are not sufficient to ensure the relevance of the information collected. The questionnaire is also a key element of your study.
help Find out more about writing the questionnaire