Sourcing and Data Quality
From sourcing through final data cleaning and checks, Feedback Loop ensures quality data to drive your critical decisions.
Sourcing
Feedback Loop sources participants from the Cint and Lucid marketplaces, giving our clients access to the world’s largest community of survey research participants. Both have quality control measures in place at the aggregate and individual survey levels. These measures include fraud detection, geofencing, and deduplication checks.
These marketplaces independently evaluate their suppliers for quality of responses, accepted completes, and consistency of quality over time. This helps ensure that participants represent your target audience, that they are engaged with surveys they take, provide accurate responses to questions, and only take your survey once.
Ensuring Quality of Responses
While participants are matched to surveys based on demographics and custom criteria, participants may have other characteristics that can influence data quality – such as how the topics of your survey and the phrasing of questions interact with their values and preferences. We can mitigate the potential impact of these interactions by diversifying the sample we collect.
Marketplaces first profile suppliers to understand the behaviors and attitudes associated with certain panels, or groups of participants. Knowing the demographic, behavioral, and value composition of panels allow researchers to optimally mix suppliers to promote diversity and ensure the representation of the general population.
Acceptance
Percentage of surveys accepted is another measure of response quality. To be accepted, a survey must not only be completed but must meet the standard of the buyer (such as Feedback Loop). Acceptance rates for suppliers are compared against the marketplace as a whole to eliminate low-quality suppliers from the pool.
Cleaning
The final step in the data quality process involves data cleaning. Feedback Loop’s research platform itself has built-in proprietary AI-powered tools that are continually learning to improve the filtering of low-quality responses.
Consistency
Measures of consistency evaluate the preferences and opinions of each supplier’s respondents on a quarterly basis to help ensure that the quality and types of participants are not changing drastically over time in ways that could affect data quality. Traits measured include brand loyalty, value-seeking behavior, technology adoption, media usage, and gaming activities.
Additional Quality Control Measures
Feedback Loop also implements additional quality control measures, including the use of a third-party fraud protection provider, Research Defender, whose sole focus is monitoring data integrity. Research Defender works by assessing the ways participants are engaging with surveys. This enables the detection of participants who are not providing reliable data, as well as bots and other forms of fraudulent input.
