Occam offers warm mailing model for charities
Database solutions provider Occam has launched occam.predict, a predictive warm mailing model designed to assist fundraisers achieve more sophisticated targeting on warm appeals. It identifies which characteristics are most likely to be found amongst responders and these characteristics are used to predict those who are most likely to respond.
Occam says that use of its bespoke predictive model to target warm appeals “increases response rates and ROI as well as reducing costs and mail wastage.”
The company explains that, on average, most charities mail their supporters five to six warm appeals each year, using a typical recency/frequency/value (RFV) segmentation. Response rates overall will be between 12-15%: therefore, around 85% are not responding. The ideal model, say Occam, would be to increase response rates but mail fewer donors, yet achieve a higher response rate and ROI.
Advertisement
Occam.predict works by scoring all individuals on the database as to their likelihood to respond. Hundreds of variables are used to determine the individual’s score. This enables selection and mailing of those donors who are more likely to respond, whilst at the same time dropping those who are less likely to respond, yet still providing volumes large enough to achieve economies of scale.
Occam.predict does not need to change a charity’s existing mailing segmentation. Instead, it allows for targeting within these segments so that donors can be ‘hand picked’ rather than blanket mailing whole segments. Scoring takes place at an individual level within segments (for example, lapsed or enquirers) that would not have been mailed because of previous poor results for the segment as a whole.
Scott Logie, Managing Director at Occam, said: “Charities can now truly engage with their supporters on an on-going basis at a level appropriate to them individually.”
Occam say that the cost of Occam.predict is available “on application”. It takes approximately three weeks to build and deliver the product for a client, and they add that one day is required to run the model and score data before each appeal.