MailingListsUSA uses a rule based system that applies computer rules for first names, surnames, surname prefixes and suffixes, and geographic criteria in a specific order to identify the ethnicity, religion, and language preference of an individual.
Unlike systems using either exact spelling of surnames, or matching against nationwide lists pre-coded with surnames, ML-USA uses a comprehensive analysis process that results in both a higher match rate percentage and a higher degree of accuracy, as well as a broader and more precise breakdown of ethnicity / religion / language preference classifications. Competitors may use approaches similar to ML-USA's or they may use simple geo-coding systems, but their results are nowhere near as robust.
We identify language preference by first identifying unique first names. For example, ML-USA will say a woman whose first name is Marisol probably speaks and thinks in Spanish whether or not her last name is Lopez or Koslowski.
Second, ML-USA examines identified surnames with commonly used first names for each surname's ethnicity to predict language preference. For example, the surname Garcia is most often Hispanic while the first name Pablo is common but not unique to Hispanic. ML-USA will predict that Pablo Garcia speaks Spanish. However, if Pablo's last name were Debrito, which ML-USA identifies as Portuguese, we would predict that this Pablo is Portuguese and speaks Portuguese.
Demographic data suggest that direct mail is the only effective marketing medium to access this lucrative market.