The move to 1:1 marketing has paid off for Haven Life, which was purchased by Mass Mutual a decade ago and launched in 2015. Matt Myers, Head of Customer Acquisition, told Friday's Data & Programming Summit it has resulted in a 78% improvement in CAC and a 44% improvement in applicant quality and likelihood to purchase or its propensity score.

After several false starts, including a bit of spray and pray, Haven Life over the last year built a 1:1-person based marketing model with 200+ model attributes, using underwriting decisions and a real-time feedback loop to its ad partners, and found it was targeting potential customers correctly. Those customers would be those less likely to die over the next 30 years, discovered via fitness data, type of car driven, etc. It lowered CAC from 52.7% in 2015 to 47.5% in 2018.

Haven Life's model works by defining customer segments, adding 400+ data attributes to each customer record, selecting attributes with the largest deviation from the population, stack rank prospecting set using those attribute model scores, activating audience across platforms, and using a champion vs challenger mentality.

Key takeaways:

Start with the correct seed audience.

Choose the right variablees.

Religiously test new data sources.

Bring a champion vs. challenger mentality to your models.