This post was originally published here (American News Report)
Remember when selecting your prospect or donor universe was straightforward? Remember when a program had one or two models to supplement straight RFM segmentation? Those days seem far in the rear-view mirror. Today’s programs use multiple models for multiple uses across multiple channels of your fundraising program.
From renewal and lapsed audience segmentation, prospect universe selection, sustainer recruitment, telemarketing calling file selections and merge optimization strategies, each model has a unique purpose. But how do you know when that model is performing at its best? How do you ensure the gains outweigh the time and cost to manage all these models?
Managing Multiple Models: Worth the Effort?
The time we spend on managing multiple models can be daunting. Keeping track of file updates, managing test quantities, updating suppression strategies, working with multiple modeling firms and ensuring the service bureau and database provider are all following the same plan is overwhelming.
Early gains make it easy for most organizations to begin incorporating more models over time. After all, models can find your organization new prospects and identify the best lapsed donors for recapture. They enable your telemarketing program to identify the best calling universes, making the program more efficient and reducing the dreaded number of caller complaints. Models have allowed acquisition programs to be more efficient by eliminating poorer performing audiences and replacing them with better prospects and allowing for tighter package and gift ask array strategies. Participation in cooperative or prospect databases have changed the way list audiences are sourced. Fundraising is more efficient in many ways, targeting the donors most likely to support your mission, so all these models are good, right?
The short answer is a resounding yes! The struggle is not “if” we should use more models, but “how” to manage the number of models we currently have and add more to the mix. When the program has models on top of models, how do you find even greater gains? Is there room for testing new models? How do you know if current models are continuing to perform as they should?
Keys to Success: Roadmap Development & New Methodologies
One key to managing it all is to develop a roadmap. A straightforward document that outlines the models in use and when they were implemented. This gives staff a method for tracking when a model should be reviewed more closely. It’s important to track these dates so the question of performance is proactive rather than reactive if results show a decline. A good roadmap helps you know when to rebuild and test a model against a control. A roadmap also allows for a macro-level view of model strategies in the overall program and should go across channels.
When setting annual goals for fundraising, it’s helpful to build in dedicated testing for model development. Work with staff and partners to discuss and review model results year over year. Use that time to discuss any shifts in strategies through the year. Is there a new channel being used? Are there new data inputs related to frequency of giving, gift size or direct mail package type or age of prospect being targeted? This information is crucial when building new models as they highlight new elements that can inform the current models. Conduct an annual review to determine whether to move away from underperforming or outdated models. This approach will help you find the strongest donors. An annual review also ensures open communication with your partners, enabling them to provide the right model for specific needs.
It is equally important for you to understand new methodologies. Models can now be developed that are a consolidation of your current models and weighted in such a way they find new prospects a single model might miss. There are new products that use online data in addition to offline purchase and transactional data. As the use of big data grows, things can change quickly. Work with your fundraising team, and understand potential gains from these new products to get the most out of your modeling performance.