Scientific Philanthropy Root-Cause Method
Use evidence to find and address the underlying driver of a social problem
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 6
- Confidence
- 97%
Scientific philanthropy treats social intervention as an evidence-led process rather than a sequence of charitable disbursements. The practitioner begins with a visible challenge, uses data and research to discover its underlying drivers, and then directs capital toward those drivers instead of merely easing the resulting pain. The proposed solution is tested in practice, monitored for evidence of impact, and refined when its assumptions prove incomplete. Once the intervention demonstrates that it works, organizations with greater capital, authority, or delivery capacity can take it forward. The mechanism converts philanthropic money into learning and proof, allowing a relatively small pool of risk-tolerant capital to unlock larger and more durable change.
Origin
The Rockefeller Foundation traces this approach to John D. Rockefeller's early idea of scientific philanthropy. Saadia Madsbjerg describes it as one of the foundation's enduring principles across more than a century of work.
Core principles
- 01Investigate causes before funding solutions
- 02Use data and research to test assumptions
- 03Address underlying drivers rather than visible symptoms
- 04Prove what works before inviting others to scale it
How to run it
- 1
Define the visible challenge
Describe the harmful outcome, the population affected, and the conditions in which it occurs without presuming a cause.
Pro tip Separate observations from assumptions at the outset.
Watch out A broad mission statement is not a sufficiently precise problem definition.
- 2
Investigate underlying drivers
Use data, research, and local knowledge to determine where the problem comes from and which mechanisms sustain it.
Pro tip Look for leverage points that influence several symptoms at once.
Watch out Do not mistake the most visible symptom for the fundamental issue.
- 3
Design a causal intervention
Choose an intervention that directly changes a supported underlying driver and state the expected causal chain.
Pro tip Write down what must be true for the intervention to work.
Watch out Avoid selecting a familiar program merely because it is easy to fund.
- 4
Test the solution
Deploy the intervention at a scale that can generate credible evidence while limiting the consequences of failure.
Pro tip Treat philanthropic money as risk capital for trying unproven solutions.
Watch out Testing without predefined outcomes can produce activity without useful learning.
- 5
Evaluate and refine
Compare actual outcomes with the intended causal chain and revise the diagnosis or intervention when evidence contradicts it.
Pro tip Use weak results to improve the model rather than defend the original plan.
Watch out Do not confuse money distributed or people reached with lasting impact.
- 6
Enable broader adoption
When evidence is strong, bring in public, private, or civic partners capable of sustaining and scaling the solution.
Pro tip Package both the evidence and an economically sustainable delivery model.
Watch out A successful pilot may still fail at scale if incentives and operating conditions change.
In the wild
Instead of treating hunger only as a shortage of agricultural production, a program examines where harvested crops disappear before reaching buyers. It identifies rapid spoilage, distant processing, weak market links, and missing purchase agreements as underlying drivers, then tests local processing, technology-enabled market links, and bulk-purchase contracts.
→ More existing food reaches consumers while small farmers retain more income from each harvest.
A foundation finds that repeated student absence is associated less with motivation than with unreliable transport and untreated health conditions. It pilots coordinated transport support and school-based health screening rather than another attendance-awareness campaign.
→ The intervention tests whether changing the diagnosed causes improves attendance more effectively than messaging alone.
Common mistakes
Funding the symptom
Relieving immediate pain can be necessary, but treating it as the complete solution leaves the mechanism producing that pain intact.
Starting with a favored solution
Selecting a program before researching the problem encourages evidence to be fitted around an existing preference.
Scaling before proving
Expanding an intervention without credible outcome evidence can amplify cost and failure rather than impact.
Is it for you?
Best for
Foundations, governments, nonprofits, and impact investors tackling complex and persistent social challenges.
Not ideal for
Immediate disaster relief where urgent humanitarian assistance must precede extensive diagnosis.
From the transcript
“it's about really understanding what the fundamental issue is.”
“to use the data and to use the research to understand what the problem really is, where the problem really comes from.”
“you're actually addressing the root causes of it.”
From the episode
Interview With Saadia Madsbjerg: Masters in Business (Audio)
Saadia Madsbjerg