Models developed with small samples of survey respondents can provide valuable insights for tactics and strategy. To be really useful, though, we need to be able to apply them to broad groups of interest. To do so, enlist one of our several sources who can apprehend 2,000 or more variables to our survey respondents and use that data to re-estimate our model and arrive at a new, generalized model that can be applied to any database.
SPH develops a predictive model based on rich survey data from a small sample of target individuals and sends the sample to a third-party data appending firm, such as Acxiom, KBM or Experian.
The third-party appender adds 2,000+ other variables available for every household in the U.S. and returns the data file to SPH with the additional data.
Our team of experts use the expanded file and the added variables to re-estimate our model, tweaking it until we get an acceptable level of predictability and can apply the new model.