The Learning Agreement: A Better Contract Between a Family and Its Philanthropy

The decision hidden inside a grant

A grant can purchase activity, support an institution or test a proposition. Those are different purposes, yet they are often compressed into a single promise to demonstrate impact. A family office should identify which purpose it is funding before asking a grantee to produce a dashboard. Otherwise the organization may report the most countable activity while the family assumes that a difficult social problem has been solved.

This article proposes a learning agreement: a short, jointly written record connecting the grant to an important uncertainty and a future decision. It is an original editorial proposal, not a tested intervention or a replacement for the grant's legal terms. The purpose is to make honest revision possible before a donor's reputation becomes attached to a preferred conclusion.

Use evaluation to ask, not to decorate

The OECD's evaluation framework distinguishes relevance, coherence, effectiveness, efficiency, impact and sustainability. It also emphasizes thoughtful application rather than treating every criterion as a mechanical requirement.[1] That provides a useful vocabulary, but not a universal questionnaire for every gift. A local food program and a long-term scientific fellowship need different evidence, time horizons and opportunities for participant involvement.

A separate OECD guide describes the tension between an organization's own learning needs and external demands for accountability, including the capacity required for meaningful measurement.[2] The editorial implication is practical: fund the work of learning. A reporting requirement without budget or staff time is still a demand on the grantee's resources.

Write the agreement around a real choice

Before approval, the foundation's program lead and the grantee should name one decision that evidence can change. Examples include extending evening opening hours, changing an eligibility process or ending an ineffective outreach channel. Record the population, proposed mechanism, expected timing, foreseeable harms and credible alternative explanations. Ask affected people which outcomes matter and how feedback can be offered without threatening their access to support.

Specify the minimum record needed to make that decision. Attendance may establish delivery; it does not by itself establish improved wellbeing. A short interview may reveal an exclusion problem that a large attendance total conceals. Choose a small combination of administrative records, participant feedback and outcome observations, with consent and data minimization appropriate to the context. Do not collect sensitive personal details simply because a donor might later find them interesting.

A hypothetical pilot

Imagine a family funding a community reading program. The first question is whether the program reaches children who cannot attend existing sessions. The learning agreement asks the provider to compare access barriers across two proposed schedules, record repeat participation and collect voluntary family feedback. It does not promise that eight weeks of attendance will establish an effect on graduation or lifetime income.

The program lead budgets staff time for the exercise. A community representative helps interpret the findings. The family's review meeting can then decide whether to extend the schedule, change transport arrangements or stop the pilot. If children with the greatest access barriers remain absent, a rising participant count is insufficient grounds to declare success. The case is illustrative; it reports no actual grant or measured effect.

Govern the response to unwelcome evidence

The foundation director owns the decision calendar; the grantee owns operational records; a named evaluation adviser checks whether conclusions exceed the design. The agreement should distinguish an operational failure, a weak theory and an inconclusive observation. These call for different responses. Automatically withdrawing support whenever results disappoint can reward selective reporting rather than good practice.

Track whether agreed decisions actually occurred, the proportion of reporting requirements used in those decisions, the cost of collecting evidence and the time taken to respond to participant concerns. Pause collection if privacy protections fail or the burden begins to undermine delivery. Stop or redesign an activity when credible evidence shows harm, when the mechanism no longer makes sense or when the agreed decision cannot be informed by the proposed data. Document the reason rather than quietly changing the original objective.

What the family should keep

Retain the original question, versions of the agreement, a concise evidence summary, dissenting interpretations and the final decision. Share an appropriately anonymized account of what changed with the grantee and participants. Families can compare learning across grants without ranking unlike causes by a single score.

This approach does not establish causal attribution or determine the best philanthropic vehicle. Those questions require context-specific expertise. Its narrower ambition is to turn generosity into a relationship in which evidence can alter action, and in which candor is treated as an asset rather than a threat.

Editorial method, evidence and limitations

This original AI-assisted editorial analysis proposes a decision framework. It is not a new empirical study, systematic review, or independently human-peer-reviewed article. Cited institutional sources inform the framing; the workflow, illustrative scenarios and suggested measures are editorial proposals. No source author, conference organizer or sponsor is represented as endorsing the article.

Original AI-assisted editorial framework; no causal evaluation, donor survey or validated effectiveness claim. Evaluation needs and legal grant terms vary by context.

Open the cited sources

1. [1] OECD (2021), Applying Evaluation Criteria Thoughtfully ↗
2. [2] OECD (2023), Policy Guide on Social Impact Measurement for the Social and Solidarity Economy ↗