Foundation Portfolio Analysis: What Your Grantmaking Adds Up To

Foundations make grants one at a time. Each decision gets attention: an application is read, a program officer forms a view, a committee or board signs off. The process is usually careful and the individual decisions are usually defensible.

The portfolio those decisions add up to is a different matter. Nobody decides it. It emerges — from who applied, who was already known, which relationships were easiest to renew, and which program officer happened to carry which docket. Ask a foundation what its strategy is and you will get a clear answer. Ask what its last five years of grants actually did and the answer is usually vaguer, because almost nobody looks.

Portfolio analysis is the practice of looking. It is not evaluation and it is not impact measurement. It is the more basic question of what the pattern is, asked of your own data, before anyone asks whether the pattern is right.

Why portfolio-level questions go unasked

Three structural reasons, none of them anyone's fault.

The unit of work is the grant. Every process, meeting, and system is organized around individual decisions. Nothing in the operating rhythm surfaces the aggregate.

The data is usually not in a shape that supports it. Grant records spread across spreadsheets, board minutes, and email cannot be queried. Answering "what share of our dollars went to organizations under $500k in revenue?" becomes a week of work, so it does not get asked.

Aggregate answers can be uncomfortable. A portfolio view sometimes shows that the strategy on the website is not the strategy in the ledger. That is a valuable finding and an awkward one, and awkward findings need a culture that treats them as information.

The result is a genuine asymmetry: external analysts routinely compute portfolio-level facts about foundations from public filings that the foundations themselves have never computed internally.

The seven questions worth asking

Portfolio analysis does not require a framework. It requires seven questions asked honestly of your own grant records.

1. Concentration. What share of dollars went to your top five grantees? Top twenty? Boards typically underestimate this substantially. High concentration is not automatically wrong — deep support for a few organizations is a legitimate strategy — but it should be a choice rather than a discovery.

2. Cause mix. What proportion of dollars went to each area of work, and does that match the split you describe publicly? Drift here is extremely common and almost always gradual.

3. Geography. Where did the money land, at county or neighborhood level? Foundations with a stated regional focus frequently find dollars clustering in a small part of that region — usually the part where the well-networked organizations are.

4. Relationship length. What is the median number of years you have funded your current grantees? A portfolio of long relationships is stable; one of short ones is churning. Neither is inherently better, but the number should be intentional.

5. New-grantee rate. What share of grants went to organizations you had never funded? This is the single best measure of whether your pipeline is open or closed.

6. Grant size distribution. Not the average — the distribution. A portfolio of many small grants imposes application and reporting burden on grantees that a portfolio of fewer, larger grants does not. The average conceals this; the histogram does not.

7. Organization size. How large are the organizations you fund? Many foundations that intend to support grassroots work find their dollars flowing predominantly to well-established institutions, simply because those organizations apply more effectively.

Reading the answers without jumping to conclusions

Each question has a common misreading, and the discipline of separating observation from judgment is what makes the exercise useful rather than demoralizing.

The right-hand column is where the value is. Portfolio analysis is a generator of good questions, not a scorecard. A foundation that concludes "we are bad" from a portfolio review has misused it; a foundation that emerges with four specific questions for its next board meeting has used it correctly.

The external layer: what the portfolio looks like from outside

Everything above can be computed from your own records. A second layer requires the rest of the sector, because it describes your portfolio's position relative to everyone else's grantmaking.

• Who else funds your grantees. For each organization you support, which other institutional funders back it — and how much overlap does your portfolio have with any given peer?
• Where you are alone. Grantees for whom you are the only institutional funder. This is a concentration risk for them and a strategic fact for you.
• What you are missing. Organizations closely comparable to your grantees, in your geography and cause area, that you have never funded. Often the most immediately actionable output of any portfolio review.
• Whether you arrive first or later. Across your portfolio, how often your money preceded other institutional funding.
• What your grantees also receive. Federal and state government funding flowing to the organizations you support, which frequently dwarfs foundation dollars and changes how you read your own contribution.

These require the full national picture. Plinth's public dataset covers 17,896,418 grants across 205,036 grantmakers for fiscal years 2017 to 2025, plus federal spending and state checkbook data matched to recipient organizations — which is what makes the last item computable at all. Any funder page is free to search without an account.

Public filings lag 12 to 24 months, so this external layer describes the recent past. For questions about pattern and position, that is generally fine. For questions about what is happening right now, it is not.

How portfolios drift without anyone deciding

Drift is the normal state, not a failure. It is worth understanding the mechanisms, because each has a different remedy.

Renewal gravity. Renewing an existing grant is easier than assessing a new one — the relationship exists, the reporting is familiar, the risk is known. Over five years, a portfolio with no explicit new-grantee target will close almost automatically. The remedy is a stated target, even a modest one.

Application-shaped strategy. You can only fund who applies. If your application process is demanding, the organizations that complete it are disproportionately those with development staff — which means larger, more established organizations. Your portfolio then reflects your form design rather than your priorities. The remedy is to look at who starts applications versus who finishes.

Program officer portfolios. Where individual staff carry dockets, the aggregate portfolio is the sum of several personal ones. Each may be internally coherent and the total may have no shape at all. The remedy is periodic cross-docket review.

Board relationships. Grants originating from trustee connections are often the largest and longest-running, and they accumulate. This is not improper — trustee knowledge is a legitimate source of good grantees — but it should be visible in the portfolio view rather than invisible.

Success bias. Organizations that report well get renewed. Reporting quality correlates with organizational capacity, which correlates with size. Over time this quietly shifts a portfolio toward larger grantees, entirely through mechanisms that feel like good stewardship at each individual decision.

Naming these makes the portfolio review less uncomfortable. The gap between stated strategy and actual distribution usually has a structural cause, not a culprit.

Doing this as an annual practice

Portfolio analysis works best as a scheduled discipline rather than a one-time consulting engagement.

1. Pick a window. Three to five years. One year is too noisy; ten washes out change.
2. Clean the grantee list first. Deduplicate names, attach EINs. Most bad portfolio analysis is bad because "Riverside Youth Center" and "Riverside Youth Ctr." were counted as two organizations.
3. Run the seven questions. Produce numbers before producing narrative.
4. Compare against your stated strategy. Not against peers yet — against your own documents. Where do they diverge?
5. Add the external layer. Co-funders, sole-funder list, look-alike organizations you have missed.
6. Take it to the board as questions. Five observations, five questions, no verdicts.
7. Write down what you decide. Including decisions to change nothing — a deliberate choice to remain concentrated is a strategy, whereas an unexamined one is drift.

Done once, this is a project. Done annually against clean data, it takes an afternoon and becomes the most useful hour in the governance calendar.

Presenting portfolio findings to a board

The analysis is the easy part. Getting a board to engage with it productively — rather than defensively, or not at all — takes some care.

Lead with the distribution, not the headline. "Our top five grantees received 62% of dollars" invites argument about whether 62% is good. Showing the full distribution, with every grantee ranked by dollars, invites the board to look at it. Boards reason better from shapes than from summary statistics.

Bring the strategy document into the room. The comparison that matters is between the portfolio and your own stated priorities, not between the portfolio and an abstract ideal. Put the two side by side and let the gaps speak.

Anticipate the three defensive responses. They are predictable: "the data must be wrong," "that grant is a special case," and "this is how it has always been." The first is answerable with source links to filings and internal records. The second is usually true and usually applies to two or three grants, not to the pattern. The third is the finding, not an objection to it.

Separate the decisions that need making from the ones that do not. Most portfolio findings warrant no action. Identifying the two or three that do — and saying explicitly that the rest are fine — keeps the discussion from sprawling.

Repeat it annually with the same metrics. The first review generates surprise; the value compounds when the same seven numbers appear each year and the board can see direction. Changing the metrics each year destroys that.

One further note on framing. A portfolio review is an examination of the foundation's own decisions, not an assessment of its grantees. If the discussion drifts toward which organizations are performing, the exercise has turned into something else — something legitimate, but requiring different evidence and a different conversation.

Where software fits

The binding constraint on portfolio analysis is almost never analytical sophistication. It is data structure. If grant purpose, geography, grantee identity, and amount are captured consistently at the point of decision, every question above is a query. If they are not, each one is a research project — which is why they go unasked.

Tools like Plinth capture those fields as part of the normal grant workflow, so the portfolio view is always current rather than reconstructed annually. Portfolio insights surfaces what is emerging across grantee reporting, where the money goes, and how the portfolio is distributed across a region, working from the reports and applications already in the system. Because it reads grantee reporting alongside grant records, patterns that recur across multiple grantees — the ones no single report would surface — become visible without anyone reading forty documents.

Two deliberate limits are worth noting, because they matter for how this kind of analysis should be used: it does not rank or rate grantees, and it does not substitute its summary for what a program officer actually wrote. Portfolio analysis is a lens on your own decisions, not a verdict on the organizations you fund.

For foundations whose grant records are currently spread across spreadsheets, the practical first step is consolidation rather than analysis — covered in Grant Management for Large Trusts and Foundations.

Frequently asked questions

What is the difference between portfolio analysis and impact evaluation?

Portfolio analysis is descriptive: it establishes what your grantmaking pattern is. Impact evaluation is causal: it asks whether that grantmaking produced outcomes. Portfolio analysis is far cheaper, faster, and a sensible prerequisite.

How many years of data do we need?

Three to five. Fewer than three makes patterns indistinguishable from noise; more than about ten obscures recent strategic change.

What is the most common surprise?

Concentration. Boards and staff routinely underestimate the share of dollars going to their largest grantees, and the gap between the guess and the number is usually large enough to change the conversation.

Can we do this without specialist software?

Yes, if your grant records are already clean and in one place. A spreadsheet with grantee EIN, amount, date, purpose, and location supports all seven questions. The difficulty is almost always getting to that spreadsheet.

Should we compare our portfolio to other foundations?

After analyzing your own, not before. Understanding your own pattern first prevents the common error of adopting a peer's shape without knowing why yours differs. See How to Benchmark Your Foundation.

Does portfolio analysis mean we should diversify?

Not necessarily. Concentration is a legitimate strategy and so is breadth. The purpose is to ensure the shape of the portfolio reflects a decision rather than an accumulation of defaults.

How do we find organizations we should be funding but are not?

By comparing your grantees against the wider population of organizations doing similar work in your geography. This requires data beyond your own records — the look-alike analysis described in the external layer above.

Recommended next pages

• How to Benchmark Your Foundation — comparing your portfolio to peers
• Co-Funder Networks — who else funds your grantees
• What Your Form 990-PF Reveals — the public version of your portfolio
• Align Grants With Strategy — closing the gap between stated and actual
• Best Impact Measurement Tools for Funders — the evaluation layer above analysis

Last updated: August 2026