Trackmob: BI and AI products for 100+ non-profits

Context

Non-profit fundraising runs on donor relationships, and the organisations that depend on it (Greenpeace, Doctors Without Borders, the ICRC, ActionAid) have the least tolerance for a tool that costs more attention than it returns.

Problem

Acquisition, engagement and analytics were three products serving one operation, and the analytics half is where a fundraising team either finds the donor who is about to lapse or does not.

Approach

I was accountable for sustaining that line as one operation rather than three tools: what a fundraising team could act on, in the time it actually had, given rules about recurring donations and volunteer capacity that no requirements document surfaces on its own.

Non-obvious decisions

• Design the analytics for the fundraiser, not for the analyst. The person who acts on a churn signal is a campaign coordinator with an afternoon, not a data team with a quarter.
• Run discovery inside the customer's operation. Non-profit fundraising has constraints (recurring donation rules, volunteer capacity, donor privacy) that only show up when you sit with the people doing the work.
• Treat the three products as one operation. Acquisition that ignores engagement produces donors who lapse, which the analytics then reports as a problem the acquisition product created.

Outcome

The BI and AI line served Greenpeace, Doctors Without Borders, the ICRC, ActionAid and 100+ other non-profits, with acquisition, engagement and analytics sustained as one operation rather than as three separate tools.