Funders publish successes far more often than mistakes, which makes the field harder to learn from. This note records design decisions we changed after they failed in practice.
Key points
- Our early application process cost applicants more than it told us.
- Short grants produced weaker work and higher staff turnover.
- Restricting overheads pushed costs into the shadows.
- Panels without lived experience missed practical objections.
We asked for too much, too early
Our first application form demanded detailed budgets and full documentation from every applicant, including those we were never likely to fund. It consumed enormous unpaid effort in small organizations.
We moved to a short first stage that asks only what is needed to decide whether to continue, and requests full documentation only from shortlisted applicants.
Grants were too short
Twelve-month grants meant that recruitment, set-up and evaluation consumed most of the period. Teams reapplied while delivering, and staff left for stabler roles. We now default to longer commitments where the work is stable enough to justify it.
Capping overheads was counterproductive
A low overhead cap did not reduce true costs; it moved them out of sight and weakened the organizations we funded. We now expect overheads to be presented openly and fund a fair share of them.
Our review panels were too narrow
Early panels were weighted towards technical specialists. They were good at assessing method and poor at spotting proposals that would not survive contact with the community described. We changed panel composition to include people with direct delivery and lived experience.
What we are still unsure about
We do not yet have a good answer to how much unsuccessful applicant time our process consumes overall, or how to give useful feedback at volume without slowing decisions. We will publish what we learn.