An arts organization gets to know its event attendees
The Challenge
A major arts organization had a growing file of constituents gained through ticket sales to their many events. They had very little information about who these constituents were, aside from the transaction details provided in order to purchase a ticket. With a desire to serve a diverse audience, the organization did not have the information necessary to determine how well their mission objective was being met.
​
​
The Solution
A system was established to purchase selected demographic data from a 3rd party marketing source and and insert it into the customer records in the CRM system as searchable attributes. The existing analysis tools were able to report and visualize the demographic makeup of ticket buyers at various categories of event.
Using these analytic tools, staff were able to analyze where the demographics did not match the desired pattern, and adjust programming to address the gaps.
​
​
The Result
The nonprofit was able to clearly see how well they were fulfilling a key mission objective, identify issues and adjust their approach. Over time they could see, and report to stakeholders, the progress toward their goal.
A non profit association uses data to grow membership
The Challenge
A non profit association was experiencing flat membership and wanted to use a data informed strategy to grow the number of members
​
The Solution
Loblolly Technologies was engaged to examine their membership data, help to craft a data driven growth strategy and establish metrics to keep the strategy on track.
​
​
Examination of membership data showed that although overall membership was stable, attrition was high with many members churning after just one year of membership.
It was determined that a fairly modest improvement in retention (with acquisition remaining level) could yield an overall growth in membership of over 25% over four years.
A working hypothesis was formed:
“Members are not finding the value in their membership, leading them to cancel after one year”
​
​
To explore that hypothesis, the association defined a matrix of activities that they believed would provide value to members at each stage of their membership, categorized by the type of member.
​
This brought clarity to the intention of the association, but revealed that few of these activities were currently tracked by staff, giving them little visibility into the members’ risk of churning.
Armed with this information, Loblolly Technologies worked with the association’s Salesforce administration team to define and build a set of reports and dashboards to illuminate member activity and flag members who are at risk of churning. The association plans to use this tracking to design more customized outreach to members, helping them to maximize the benefits they derive from their membership.
​
A non profit explores machine learning in donor data
The Challenge
A non profit organization was curious about using machine learning to predict future donor behavior
​
The Solution
Nicole Weaver (Principal at Loblolly Technologies) worked with the development department to select two pilot projects for exploration. She then engaged an expert machine learning firm to develop and test the ML models and prepare a real life test for the predictions.
​
Pilot 1 - for regular donors, predict the date and amount of the next gift
Taking the pattern of giving for donors who had given 3 or more gifts, a model was built to predict the month and amount of the next gift. The predictions for each donor record were inserted back into the marketing system and made available for campaign segmentation. This resulted in the opportunity to save expense and customize the ask for each donor.
​
​
This diagram shows the pattern of prior giving for a sample donor, plus the predicted next gift in red.
​
​
Pilot 2 - for new prospects, gained through event attendance, predict the likelihood that they would become a donor
This non profit organization gained many contacts through the many events it held. Only about 8% of them became donors. A ML model was build to identify which of the new prospects were most likely to become donors.
​
​
​
​​The model was able to identify a pool of 40% of the entire file which contained 92% of the future donors. This allowed the client to reduce solicitation of 60% of the new prospects and save significantly.
​
​
