By Kelly Osterhout, Client Strategy & Insight Manager

At this year’s AdExchanger Industry Preview, Hilary Mason, data scientist at Accel, made a sharp observation about people’s behavior. According to Mason, users say they prefer to go online mainly to share news articles and pictures. But in reality, they’d rather read celebrity gossip and fool around on social media. Thus, a conflict exists when helping my clients connect their brands to online consumers – the need to present what we believe is suitable vs. the behavior that reveals our true inner preferences.

Planet Money featured a discussion about Spirit Airlines, a budget carrier known for low fares and poor customer satisfaction, which illustrated this exact point. On the program, Barbara Dingus, an airline stewardess, noted a specific kind of Spirit passenger – the “hate flier” – who despises traveling with the airline, yet repeatedly books with them. While these fliers resent feeling nickel and dimed for even the most basic of travel amenities, they are ultimately motivated by the low fares offered and therefore continue to fly Spirit. The conflict between what fliers say they prefer, like low costs, and what they actually do, like complain about the flying experience, is fundamentally the same challenge of modeling for online ad targeting. Given its inherent complexity, how can we really know what matters to a consumer?

Some modeling does bypass this conflict between stated and revealed preferences by focusing only on consumers’ internet browsing behavior. These models are complex and dynamic, but their strength in pattern finding lacks precision. It’s one of the main reasons you are served travel ads long after your vacation is over. The majority of ad technology operates on this very type of inferred modeling. And when done well, it can be quite effective. However, to gauge the true effectiveness of this inferred modeling, companies compare the performance of their targeting to the run of network. They query if their targeting outperforms when contrasted to indiscriminate, random ad serving, but there is no way for them to measure if their modes actually match up (in a privacy friendly way) to the person at the other end of the computer.

Resonate sits between ad technology, marketing, and consumer surveying. We ask consumers what influences their brand engagement, and then we look for dissonance between the stated and the revealed preferences. Our platform uses a PII-friendly process to ensure web browsing behavior ties back to the consumer values which train the computer models, accounting for the discrepancies in preferences. According to our clients, Resonate outperforms endemic sites and drives 96% of incremental traffic to their websites. This proves that our clients’ consumers had awareness of their brand but were not engaging with it until implementing our motivations-based targeting. No other ad technology company integrates the sophisticated output of 200,000 annual surveys into the machine training its targeting models to achieve such success.

Read more on the ways Resonate continues to advance its surveying.