The Power of Data in Marketing: 5 Tips for Using Predictive Modeling

While there’s been a definitive shift toward digital over the last decade in ad spending, there’s still a great deal of potential in traditional tactics — especially now that big data and predictive modeling can help marketers target their communications more precisely than ever. Here’s how Northlich recently took a targeted marketing approach to reach specific consumers for a population-health initiative in Alabama, plus five predictive modeling tips marketers can use to reach the right audience.

Northlich teamed with the nonprofit biotech institute HudsonAlpha on an initiative designed to increase awareness of hereditary cancers and identify those at risk — particularly those who might carry the BRCA1 or BRCA2 genetic mutations associated with elevated risk of breast cancer. The program offered free or reduced-cost genetic testing kits to people in a five-county Alabama region.

The year before our engagement with them, HudsonAlpha was working with a very small advertising budget and focused on outreach efforts such as PSAs, news releases, rack cards at local businesses, and partnerships with banks and community organizations. They saw good results, but wanted to direct their efforts more specifically to reach their targets.

This is where Northlich stepped in. By leveraging health and behavior data analytics, we identified an algorithm for finding targets with an elevated likelihood of carrying the genetic variables the kit tested for. Our hypertargeting approach supported a 300 percent growth in the number of test kits requested versus the previous year.

While the collateral for this campaign was a direct mail piece, a hypertargeting approach is remarkably useful for any number of mediums. In the days before the potential of predictive modeling was widely understood, a “spray-and-pray” approach was often used with direct tactics: Send your materials to all households in an area that might contain a member of your target audience, and hope for the best.

Using predictive analytics is a far better strategy. Here are five tips on using data to effectively reach the right targets with your marketing.

  1. Start by clearly defining your target and your primary purpose. A highly specific target gives you a clear destination, and a clear purpose for the campaign helps you craft the right messaging. A campaign focused on acquisition will differ from one aimed at retention or upselling.
  2. Invest in robust data resources. Hire an agency that has access to (and knows how to use) a high-quality data pool, and regularly check with your agency to ensure the data is recent, vetted, validated and clean. The more information you have access to and the better it is, the more effectively you can target consumers. A small database can yield an incomplete picture; trying to assess data without having all the information you need is likely to yield an inaccurate result. Investing in big data and in those who can interpret it pays off big.
  3. Combine relevant pieces of data to tease out differences between your target and the rest of their demographic. Minor variations in income range, number of children in the household, or years of education completed can signal differences in other indicators like health status, purchase intent or a propensity to spend — all data points you can combine and weigh to predict how likely your target is to take the action you want them to take.
  4. Track the effectiveness of your efforts. When a campaign performs particularly well or particularly poorly, it shouldn’t be a surprise. If it is, you have an opportunity to improve the way you measure how your consumers are reacting to your campaigns.
  5. Consider the decision-making cycle. For instance, making large purchasing decisions can take years, so a single marketing touchpoint isn’t likely to be the one that tips the prospect into taking the action you want her to take. Consider the place of each tactic within both the macro and micro view of the consumer, and weigh them accordingly.

As marketers lean into the future and work to improve our strategies, tactics and measurement capabilities, predictive modeling will be an increasingly useful tool. Investing some time up front to learn how to wield it will pay solid dividends.