Alzheimer’s Society

Alzheimer’s Society has a mature and highly successful legacy marketing programme. Legacy income contributes a substantial amount each year and has been increasing over the last 5 years.  This growth was recognised as partially due to its regionally based LMOs (Legacy Marketing Officers) who provide direct contact for prospective givers. Alzheimer’s have a substantial base of individuals to be mined from within their CRM platform, with a set of defined and recorded legacy relationships.

However the prevailing data selection methodology was based on supporter tenure & transactions. Legacy targeting was primarily done to long-term cash donors, regular donors, members, raffle players, and recently lapsed donors with a long history.  So the recognised opportunity was to both improve the selection methodology through identifying donors most likely to leave a legacy, but also to understand and inform the supporter journey, and prioritise who Alzheimer’s Society should communicate directly with about legacies.

Adroit Data and Insight were engaged to support the organisation to support the charity on this major strategic initiative, both to develop a set of propensity models appropriate to different audiences, and to support in the application of this back into the legacy programme, including planning & forecasting.

A legacy propensity model was deployed in and the theoretic gains that were illustrated in the model build phase (in excess of 85% correct classification) tested.   Adroit had also delivered underlying data in their Legasee, forecasting and planning tool, which had been used to determine a test strategy for the initial test mailing, as well as allowing a long term legacy scenario planning tool.

The initial use of the model generated strong results. However it was the response distribution that demonstrated the overriding model power, with 68% of all response, and 86% of pledge response (which was the outcome that the model was specifically predicting) coming from the first decile.  This was extremely strong and even higher than the theoretic gains of 67%, especially when considering this was on the net contactable file.

The model has also been used successfully on multiple occasions to plan events and is generating new converts into Alzheimer’s legacy programme. This was done by producing heat maps of where legacy best prospects are for event planning, and also for targeting local activity.

The model was partially used to drive invitations, and 74% of all attending responses falling into the top decile of model.  The events have proved highly successful, and the model showed how important it is as an amazing 96% revealed pledges came from the attendees within the top score decile, with all others in the next decile. Again showing how successful the model is in targeting pledgers.

These were particularly well received by the regional officers who plan and run these events as it provided  concrete insights to work with in the planning stage for the first time ever:

“This propensity model is already making a real difference helping me plan my year by clearly showing me where it would be best to target my marketing events and also how to be cost effective to suit an area of low support. I can also clearly identify the best place to hold my annual Stewardship event which aims to present the work of the Society as a whole.”

Estimates of forward value are made based on current residuary and pecuniary averages, a pledge conversion rate of 90%. Estimates based on these assumptions show that the value of the legacy income achieved through this strategy would be a significant gain.

Whilst the results are impressive in themselves, there are also additional factors that we believe in combination make this different to a “standard” legacy propensity model:

• As a strategic initiative, it has shifted the way data is used and valued both within the fundraising team and also at the operational level
• The model(s) have been fully integrated into Alzheimer’s database team, who were supported in implementing the code, testing and applying the scores – this was done by clever data engineering within FastStats
• Models were all built without using age variables (which can be applied as a filter)
  Since this has been supplied as dynamic code, the models will automatically rescore – Alzheimer’s have maintained a score history and use change codes to target increasing propensity (this will provide a “trending” and stewardship opportunity on tracking supporters within the legacy programme)
  Whilst underlying model factors might require re-weighting periodically, this recalibration is easy and quick (and therefore cheaper) to implement
• The targeting itself can reference Adroit’s Legasee planning tool to determine acceptable ROIs and cut-offs for activity, supplementing existing forecasting tools
• The models can be used at a bespoke “Region” level to support local planning activity, and have been well received by LMOs as thematic heat maps
• Other applications of the model into media and event planning are planned, meaning it will provide additional benefits

The client commented

“Producing heat maps for those with the highest propensity scores will really help to ensure we maximise the value of our legacy events.  We are now able to target supporters who we know are more likely to leave a legacy based on insight rather than on previous assumptions. It gives our legacy officers who plan and run campaigns and events something concrete to work with and enables us to plan legacy communications and the locations of our legacy events much better. It also means we can be more effective as an organisation in planning our customer journeys”