The Strongest Predictor of Foundation Funding Growth Is Not Funder Count. It Is Concentration.

In our previous analysis, we showed that nonprofits with more foundation funders receive disproportionately more foundation funding, and that the relationship holds even after controlling for organization size. But that was a cross-sectional snapshot. It could not tell us whether the pattern holds over time, or what drives it.

So we tested it longitudinally.

What we did

We tracked about 82,000 nonprofits across three years of IRS 990 and 990-PF tax filings from 2023 to 2025, excluding donor-advised funds (DAFs), fiscal sponsors, and pass-through foundations to isolate direct foundation relationships.

The earlier cross-sectional pattern held: gaining foundation funders was associated with stronger funding growth, and twice as many organizations gained funders as lost them over the period. But when we controlled for how concentrated an organization's foundation funding was in its single largest funder, a stronger signal emerged.

The concentration gradient

We measured each organization's top foundation funder as a percentage of total organizational revenue. The results form a smooth, monotonic gradient:

Organizations in the lowest concentration band were more than three times as likely to grow their foundation funding as those in the highest band. The crossover point, where outcomes shift from positive to negative, falls between 5% and 10% of total revenue for most organization sizes.

Notably, the median number of funders was similar across all bands (2 to 3). Organizations with the same number of funders had very different outcomes depending on how the dollars were distributed.

This is not a size, sector, or funder count effect

The gradient holds within every size band we tested, from organizations under $500K in revenue to those over $25M. It holds within the same funder count (among organizations with exactly 4-5 funders, 69% of the low-concentration group grew vs 17% of the high-concentration group). It holds across all NTEE sectors, all 50 states, and with or without DAFs included.

We validated the concentration measure against an independently-sourced IRS income classification (71% band agreement). We confirmed the gradient using both two-year and one-year measurement windows. We tested for foundation-side supply shocks and found the gradient unchanged after excluding nonprofits funded by the 50 largest declining foundations. We tested for government funding confounds and found the gradient holds for organizations both with and without significant government grants.

Regression to the mean: a real factor, but not the explanation

One meaningful caveat: 39% of organizations in the most concentrated band (35%+) had 2022 as an abnormally high year relative to 2023, suggesting some of the observed decline reflects regression to the mean rather than structural risk.

After excluding these spike-year organizations, the gradient remains monotonic and the most concentrated band still shows negative outcomes. But the magnitude decreases: the median decline for the 35%+ band drops from roughly -55% to roughly -30%. The finding is real, but the headline figures overstate the structural effect for the most concentrated segment.

What drives the difference

When we decomposed funder turnover, we found that diversified organizations (top funder under 40% of foundation funding) and concentrated ones (top funder over 60%) gain and lose funders at roughly similar rates. The difference is in the dollars.

At most size levels, the median diversified organization lost zero dollars to funder departures. Their key relationships were retained. When turnover did occur, incoming funder dollars exceeded departing funder dollars 65-74% of the time.

Concentrated organizations experienced real dollar losses to departures at every size, and incoming dollars exceeded departures only 41-66% of the time. Diversified organizations retain their key relationships while adding new ones. Concentrated organizations churn through funders at similar rates but absorb larger losses each time a relationship ends.

Among large nonprofits ($25M-$100M in revenue) with meaningful foundation dependency, the contrast is stark: diversified organizations saw median growth of +22% and $1.1M in additional funding, with 4.5% experiencing severe decline. Concentrated organizations of the same size saw median declines of -52% and losses of $4.1M, with over half experiencing severe decline.

Implications

Foundation relationships do appear to compound over time. But the compounding works in both directions. Low concentration compounds growth. High concentration compounds risk.

The data suggests that foundation diversification is less about the number of funders and more about how exposed an organization is to the departure of any single one. For development teams, this reframes the goal: the question is not only how to attract new funders, but how to build a portfolio where no single departure is severe.

---

Full methodology and robustness testing

Panel: ~82,000 nonprofits present in all three years of IRS 990 and 990-PF tax filings (tax years 2022-2024, filed 2023-2025) with at least one matched foundation funder.

Exclusions: Grants from DAFs (Fidelity Charitable, Schwab Charitable, Vanguard Charitable, etc.), fiscal sponsors, and other pass-through foundations excluded (~2,763 foundations removed) to isolate direct foundation relationships.

Concentration measure: Top funder's grant total (from foundation_grants, tax year 2022) divided by total organizational revenue (from most recent 990 filing via nonprofit_grant_portfolio). Validated against independently-sourced IRS income codes (income_cd from BMF data); 71% of organizations land in the same concentration band using either revenue source, and the gradient is near-identical.

Robustness checks (14 tests across 3 rounds):

Strictest combined filter: Excluding spike years, requiring recurring top funder, baseline >= $10K, 2+ funders: gradient is +129% to -29% (74% vs 30% grew).

Caveats:

• Survivorship bias: organizations that lost all foundation funders exit the panel. Dropout rate is highest (29%) for the most concentrated band, meaning reported outcomes understate risk.
• Regression to the mean inflates the magnitude of decline for the 35%+ band. After excluding spike-year organizations, the decline is approximately -30% rather than -55%.
• Partial mechanical correlation: concentrated organizations are mechanically more exposed to the departure of their top funder. The gradient holds even when the top funder stays, but is amplified when they leave.
• Correlation, not causation: organizations with stronger programs and leadership may both attract funders and grow funding.
• Three-year observation window: we cannot confirm whether these patterns are structural or period-specific.

Spearman rank correlation with funding % change (predictor comparison):

• Top funder / total revenue: |rho| = 0.44 (strongest)
• Funder count: |rho| = 0.21
• Top funder / foundation funding: |rho| = 0.14
• Foundation dependency %: |rho| = 0.01
• Organization size: |rho| = 0.001

Data source: SciRise Foundation Intelligence dataset