Predictive Analytics in the Not-For-Profit Sector A lot has been said about how data science can be applied to point-of-sale data to help retail businesses better understand their customers’ behaviour. Who are our best customers? What kinds of products do they buy, what regions do they live in and how frequently do they come to my store? From this understanding, marketing strategies can be crafted with one goal in mind – to increase profits.

The same principles can apply to not-for-profit groups and charities, but it’s not as simple as replacing the word “customer” with “donor”, although that is a good way to start thinking about it. Profit is no longer the goal; it may simply be awareness. In a sector where promotional budgets can be highly limited so as not to expend funds better spent on the central cause or programs that the group serves, targeted, data-driven marketing is essential.

To begin with, existing lists of donors or members can be mined, anonymously, to gain insight into the regions in which these people live. Third party socio-demographic data can be joined with the results to find other regions of similar affluence that previously may have been underexposed to marketing efforts. For charities that are worldwide in scope, such as the Red Cross, there may be vast areas of potential from this effort alone.

Groups that are more local, such as an animal shelter, may rely on donations through social media exposure or fundraising events. All campaigns can be tracked – at what times of day or on which days of the week are people most likely to respond? Are there seasonal correlations? Correlations with other keywords – such as “animal cruelty” – found on social media platforms like Twitter or Facebook? Or is there a negative sentiment about the programs being offered by your group? The effectiveness of any campaign can be easily measured through a business intelligence tool, to better inform future promotional efforts.

Not everyone is responsive to the same message. Personalized appeals to certain demographics of supporters could be highly effective. Charitable groups can measure the optimal delivery channel for receiving donations – e-mail, web site, phone, door-to-door, etc. – and segment customers by this criteria. Individual and corporate donors can be segmented into groups so that an understanding is gained as how to best solicit each group, and how to effectively reach similar, untapped donors. Even anonymous non-donors, those who simply browsed the web page of the organization, may leave a footprint in the server log files that can be turned into geographically clusterable information. Can this data be used to, at minimum, identify potential volunteers?

The last area where predictive analytics can help is in identifying churn – the donors or members who were once in the upper segment of participation but are no longer active. Have they lost interest, or is contact information simply out of date? Are there statistical correlations between churned donors – can churn be predicted before it happens or even averted? Perhaps communication frequency with the donor was too infrequent (i.e. yearly) and the cause has slipped out of mind, or worse still, communication was too frequent.

Whether the goal is profit-driven or philanthropic in nature, if data is available, analytics can leverage it.

Related Post

Leave a Comments