When it comes to understanding consumers, researchers always want to know more about their target audience. We want to know how they feel, where they shop and what their family looks like. Trying to gather these nuanced insights about an audience using just one approach is challenging and often requires compromises like low response rates, less detailed insights or outdated data.

That’s where data fusion comes into play. Data fusion was put into practice in the U.S. in the 1970’s and has since spread globally. Data fusion solves for these issues by blending multiple data sources into a single respondent-level data set.  Examples include fusing credit card purchasing data with quantitative survey data to understand purchasing behaviors and purchase drivers. Fusion produces more useful information than any one data source can provide on its own. This fused data set can be used for robust analysis across a variety of categories.

The fusion process matches and blends respondent-level data sets using common characteristics. Respondents from one survey or database are matched with respondents from another survey or database. This matching process is done based on common characteristics in the two datasets. Typically these common characteristics are demographics, geography or behaviors.

Data fusion makes the assumption that respondents with similar characteristics are likely to look, act and feel in similar ways across different categories of data. So, for example, a female respondent age 25-34 with children on one data set is a good proxy for a similar respondent on another data set.

 

While single source data is always ideal, data fusion has many benefits. Because the process utilizes existing data sources, the process is cost-effective and convenient in the absence of single source data. Fusion harnesses the best data available in a process that can be repeated and updated.  A fused database using the best measurement data available will provide an accurate, flexible and actionable data set, to optimize decision making and maximize efficiencies for researchers.