People often think of predictive analytics as something that’s primarily useful for commercial organisations or is too expensive for charities to make use of, but this really isn’t the case. At Smart Vision we work with a wide range of charitable organisations, many of whom are using SPSS Statistics or SPSS Modeler software for data mining and analytics. The vast majority of these organisations use statistical analysis for programme delivery, outcomes measurement and policy work. It’s no secret that many need to do this in order to report back to grant making bodies. This is a fairly conventional use of statistics and it’s work that tends of be done by trained statisticians rather than by marketers or fundraisers directly.
However, more recently I’ve noticed that many charities are becoming much more sophisticated users of analytic approaches to driver better fundraising as well. Increasingly, they’re using advanced analytics to better target donor segments with the greatest propensity to give; to determine the best amount to ask each segment to donate and to understand what factors affect their marketing response, attrition and legacy giving rates. There are some really clever charities out there who are exploiting the very best of the commercial world’s implementation of predictive analytics whilst being mindful of cost and the need to avoid excessively large investment.
I’ve also heard of some frightening misuse of statistical approaches over the years. For example, I know of a charity that created an ‘ask ladder’ that inadvertently downgraded their regular one-off givers by asking for 75% of their last donation each time. Another noticed a correlation between levels of giving and donor’s first names and so was intent on targeting by name, completely missing the fact that name was a proxy for age – it was actually donor age driving the trend and not the fact that the best donors were called Agnes and Betty. Fundraisers often spend a lot of time optimising the language of the ‘ask’ and the imagery of their mailers only to blast it to tens of thousands of people regardless of their propensity to respond. I saw an example recently of a charity that mailed 5,500 supporters to try to get 10 people to a major donor event, totally missing the possibility of segmenting down the invitation list to cut the cost of the mailing.
Mistakes like this generally happen when an organisation invests into predictive analytics but without also investing into making sure their staff have the necessary skills to be able to use the software effectively and, crucially, interpret the results correctly. When budgets are tight it can be very tempting to cut corners when it comes to training but that’s really a false economy and these examples show why.
I regularly work with charities who see a significant return on their investment, even taking into account the costs associated with training, support and so on. Most charities would fall over themselves for a 3-5% increase in donations and consider the investment in software and training that can make that happen to be money well spent.
At Smart Vision we have helped many charities to use predictive analytics and the data they already hold in order to make small changes to their marketing which can have big effects on their bottom line. We’re hosting a half-day seminar on 18 October 2013 designed to help charities learn more about how predictive insight can help with effective donor management.