In simple terms, unlike other insights in our report, we calculate socioeconomic data from the demographic census that represents the audience or the segment.
But, let us take a deeper look into how it’s done. Here is a more transparent explanation of how we calculate this tab, and the different factors involved in obtaining the insights.
For context, we use Facebook population as census data. From Facebook, we receive data on these socioeconomic areas:
- education level
- relationship status
- family status
- household income (only US)
When determining the socioeconomic insights, for every audience and segment, we look into 4 attributes and find the most relevant value for each one of them, in order to build a proxy of that segment: what are the country, the gender, the interest and the age that best define the segment. In order to consider the top value as the most relevant for each of those attributes, it has to surpass a certain threshold, these are as follows:
- Country: above 50%
- Gender: above 70%
- Interests: above 50%
- Age: above 25%
What happens if an attribute does not meet the threshold?
For any of: gender, interest or age, if the value is lower than the threshold, then we ignore that attribute in our creation of the proxy segment.
For countries, if there is no country above 50%, we fall back to a global audience.
So if there is no interest, gender, age, or country to be matched, the socioeconomics group will be simply global overall.
In another example, if we could only map gender and interests, it would be top gender top interest, and global as a country.
Here’s an example, if a segment’s top country has a 10% value (top of the segment but a low %, and below the threshold) we don’t apply it. If no value gets above the threshold we fall back to the “worldwide” trait. If the top country is Brazil with 10%, as it does not surpass the threshold value of 50%, we consider the country trait to be “Worldwide”.