“It was an extraordinary outcome for us. Better still is our continued improvement, especially in our primary objective of new member acquisition.”
– Channel Director, National Non-Profit Organization
Our customer is a major not-for-profit organization that advocates on behalf of its members. The organization runs a number of different programs to provide education and assistance with the everyday issues that face their target demographic.
While its membership is already one of the largest independent groups in the U.S., it is actively seeking to recruit new members on a daily basis. One of the ways it does this is through newspaper inserts and shared mail advertisements. These efforts are managed by a small group within their larger marketing organization. The program director understood the need for a more rigorous analytic solution for improving the efficiency of the program and optimizing acquisition. In addition, the requirement for maintaining fast time-to-market and improved decision making required a solution that automated much of this complex process.
Data and Approach: The nature of the newspaper and shared mail media outlets presented a significant marketing challenge — not knowing the exact households receiving the media. We utilized geo-demographic data describing the 50,000+ potential newspaper and shared mail delivery zones. In addition, we relied on historical targeting data at the zip code and delivery zone level. For example, we found that tracking cadence — an indication of how recently and frequently a delivery zone was targeted — provided significant information for improving decision making. Other data available included past zip code response performance, seasonality, and client-level demographic indices.
Predictive Analytics: PredictIQ was used to create predictive models of expected zip code and zone level response rates. Separate models were built for each sub-channel. Models are refreshed on a yearly basis in order to continually improve performance based on new data, and are continually assessed based on in-market performance. All potential delivery zones in the client’s print media network are scored each campaign since dynamic factors such as cadence and past performance are always changing.
Optimization: Making the media buying decision each campaign requires the identification of which placement zones will provide the best return for the campaign’s budget and objectives. But it’s not as simple as ranking model scores. A number of business constraints must be adhered to, including organizational diversity objectives, budget allocation across sub-channels, and newspaper minimum insert
requirements, among others. We use OptimizeIQ to easily enter these constraints in business language and generate optimal targeting decisions each campaign.
- Over 90% improvement in acquisition rates and reduced cost per acquisition
- In-market model lift is over 5x from top to bottom zip code deciles
- Complex modeling and optimizing acquisition processes are now automated and integrated