Lifetime value can improve overall donor revenue
Lifetime value calculations [LTV] are rarely undertaken as they seen as too complex to derive, often requiring the aggregation of data from more than one transaction system... all the more so in the non-profit sector where transactions for donors and members are often held separately.
Qbase were asked to calculate the actual LTV of donors to a major charity as the first stage of this project. The second stage was to develop a multi-variate propensity model which would enable the client to stream donors according to their likelihood of their attaining high, medium or low LTV.
Each supporter stream then received different communications and campaigns, in essence focussing more marketing expense on the donors likely to have the highest LTV.
The final stage of the project was to provide training for the client's internal analysts and to migrate the solution in-house.
Qbase worked with the client to identify all relevant data sets that were held within their various systems, including donor contact details, giving history, relationship types and changes over time, communications history, payment method and gift aid details. That data was then hosted at Qbase, using Microsoft SQL Server as a data warehouse and FastStats Discoverer as a business insight front-end.
The actual LTV of each donor was calculated, enabling us to segment them into High, Medium & Low LTV groups. We then used a combination of SPSS, AnswerTrees & PWE to model each segment, based on their actual previous behaviour with the client. By analysing patterns of activity, Qbase were able to identify 'triggers' which could be used as predictors of likely future behaviour, i.e. moving up or down the segments.
Qbase then tested the predictiveness of a number of third party demographic and lifestyle attributes, including date of birth, home ownership data and debt history. Having identified the most predictive attributes, these were appended to the donors and a profile produced.
Finally, on-site and off-site training of three client analysts was carried out, over a 2 month period, to ensure that the skills necessary to continue this work were migrated into the client marketing function.
The results have been a significant improvement in average donor value, generated through the ability to target investment where it is required. High LTV donors and those showing signs of becoming high value receive a greater proportion of the budget, whilst low LTV donors receive a reduced investment.
Measuring of the triggers as Key Performance Indicators has enabled the client to proactively discourage churn behaviour whilst encouraging growth behaviour.
Use of the profiles has enabled the client to target prospects who show attributes which are positively correlated with the High LTV segment.
The models are now run and re-iterated by the client's internal analysts, due to the skills transfer Qbase undertook.
If you want to correctly calculate LTV and assess how it could impact on your supporter revenue, please contact Paresh Patel