Nonprofit cases optimize impact per unrestricted dollar, with donor restrictions as hard constraints. They are not government KPI cases and not healthcare provider cases.
Last full pass: 24 August 2026.
Bridgespan-style and MBB social-impact rounds still use a case structure: clarify the objective, build a tree, do the math, recommend. The twist is the objective. It is rarely “profit.” It is meals, literacy, relapse-free months, or lives — subject to restricted grants, reserve policy, and a board that fetishizes the overhead ratio. If the client is a ministry, use the government case interview. If it is a hospital with a charity-care mission, you are still mostly on healthcare economics plus a payer mix. This page is NGOs, foundations, and mission-driven operators.
What you are deciding
- Where to put the next $X of unrestricted funds (deepen vs expand geography).
- Whether to accept a restricted grant that does not pay rent.
- Whether a program should close because cost per outcome is 3× the alternative.
- Earned-income vs donation mix (a social enterprise still has a mission constraint).
Clarify the outcome definition (and whose), the time horizon, and what money is restricted. Ask whether overhead is allowed as a % of the grant.
Exhibits you should expect
| Exhibit | Trap |
|---|---|
| Restricted vs unrestricted revenue | Restricted cash can look like a surplus you cannot spend on HQ. |
| Cost per output vs cost per outcome | Meals served ≠ children at reading level. |
| Overhead ratio | A 18% overhead can be the condition for program quality; a 8% ratio can mean starved measurement. |
| Utilization (shelter beds, clinic slots) | Empty capacity with restricted food budgets is a mismatch, not “inefficiency” in the CPG sense. |
| Donor concentration | One foundation at 40% of revenue is a going-concern risk. |
Traps that are unique here
Overhead is not the enemy. Interviewers want you to use it as a talking point and then reject it as the decision metric.
Restricted funds. A $4m meals grant that leaves $1.2m “left over” cannot pay the lease if the grant forbids it.
Attribution. Expansion into a new city may count “new kids” who would have been served by someone else. Incrementality belongs here as much as in marketing.
Practice a nonprofit impact case
Run a scored case where restricted cash and cost-per-outcome have to beat an overhead-ratio story.
Worked mini-case: leftover meals money is not a surplus
Prompt. GoodTable has a $4.0m restricted grant to serve meals. Cost per meal $3.10. They can serve 900k meals → $2.79m. $1.21m would remain, but the grant cannot pay rent or M&E. Separately, the board has $1.8m unrestricted and two proposals: (A) new-city expansion claiming 2,400 extra children at grade-level reading, (B) deepen the current city for 4,100 extra children (case inputs; treat as program-reported, not an RCT). Overhead hawks want to reject anything that raises the 18% admin ratio. What do you do?
Math. Restricted “leftover” $1.21m is not deployable to reading. Do not present a combined surplus. Unrestricted $1.8m: cost per additional child (A) = 1.8m / 2,400 = $750; (B) = 1.8m / 4,100 ≈ $439. Even before incrementality, B is more efficient. Flag that A’s 2,400 may be less incremental (new city has other providers).
On the meals grant: either scale meals toward the 4.0/3.10 ≈ 1.29m meal cap if operations can, or renegotiate with the donor for a 15% admin cost and M&E — a normal, professional move, not “waste.”
Recommendation. Put unrestricted $1.8m into deepening (B), with a simple before/after reading measure on a sample, not a new-city ribbon-cutting. On meals, ask the donor to allow rent and measurement rather than leaving $1.21m stranded. Tell the board the 18% overhead is a control, not an objective; cutting M&E to 12% would make the 4,100 figure untrustworthy. Risk: donor concentration on the meals grant; start a 12-month unrestricted pipeline. Next exhibit: restricted-fund stack and true incremental outcomes in the current city.
What a generic profitability tree misses here
- Revenue is not fungible. Restricted vs unrestricted is the first split, not price vs volume.
- The unit is outcomes, then outputs, then cost.
- Overhead ratio is a political metric, often inversely related to learning.
- Mission constraint. You cannot “exit” the high-cost last-mile population if that is the mission — you can still compare programs inside it.
- Volunteers and in-kind distort unit cost if you ignore them or double-count them.
If you answer GoodTable like a retailer, you will call $1.21m a profit pool. If you answer it like a ministry, you will optimize a published KPI the donor did not fund.
See where you stand on a nonprofit case
Practice restricted-fund logic and cost-per-outcome math.
Related guides
- Government and public sector case interview
- Marketing case interview (incrementality)
- Profitability framework
CoachNed is independent and unaffiliated with McKinsey, BCG, Bain, or other firms named for interview-style practice. Cases on CoachNed are AI-simulated.