Skip to content

Methodology

How the peer benchmark is built.

Every number in the Nonprofit Peer Benchmark Lookup comes from public IRS Form 990 filings. This page lists the lines used, how peers are chosen, how percentiles are placed, and where the data cannot be compared. Edition built 2026-09-30, covering 173,008 organizations.

1. Data sources

All sources are free and public. Nothing is surveyed, estimated, or entered by hand, and no private data from any organization is used.

FiguresForm 990 linesRelease
Contributions and grants, program service revenue, total revenue, total fundraising expenses, total expenses, net assetsForm 990 Part I, lines 8, 9, 12, 16b, 18, and 22, current and prior yearIRS e-file release, NCCS efile table F9-P01-T00-SUMMARY, tax years 2021 to 2024
Program, management, and fundraising expense splitForm 990 Part IX, line 25, columns A to DNCCS efile table F9-P09-T00-EXPENSES, tax years 2022 to 2024
City, state, website, 501(c)(3) status, year of formationForm 990 headerNCCS efile table F9-P00-T00-HEADER, tax years 2022 to 2024
Mission category (NTEE code) and subsectionIRS Exempt Organizations Business Master Fileirs.gov/pub/irs-soi/eo1 to eo4, monthly release

The IRS e-file releases are read as published by the National Center for Charitable Statistics at the Urban Institute (efile v2.1 tables). Names are shown as filed, in title case.

2. Which organizations are included

Looked up

  • Organizations that filed the full Form 990 electronically for tax year 2022 or later. Nearly every 990 is e-filed now, so this is close to every full-990 filer.
  • Their three most recent tax years, where available, for the trend figures.

Placed against peers

  • 501(c)(3) organizations (the 990 header box or the Business Master File subsection).
  • Total revenue from $250,000 to $250 million on the newest filing.
  • A full-year return (350 days or more) that is not a group return.
  • Cohort distributions use organizations whose newest filing is tax year 2023 and 2024.

Not covered: 990-EZ and 990-PF filers, organizations that do not file (most churches), and organizations whose newest public filing is older than tax year 2022. If a search finds nothing, the lookup offers a hand-run benchmark.

3. How peers are chosen

A peer cohort is every placed organization in the same NTEE major group and the same revenue band. Nothing else enters the definition: not geography, not age, not staff size.

Revenue bands

  • Under $1M (from $250K)
  • $1M to $3M
  • $3M to $10M
  • $10M to $15M
  • $15M to $22M
  • $22M to $30M
  • $30M to $250M

The bands between $3M and $30M match the ones used across this site, so a benchmark and a working session describe the same organization.

NTEE major groups

  • A Arts, culture, and humanities
  • B Education
  • C Environment
  • D Animal welfare
  • E Health care
  • F Mental health and crisis intervention
  • G Disease and disorder organizations
  • H Medical research
  • I Crime and legal services
  • J Employment and workforce
  • K Food, agriculture, and nutrition
  • L Housing and shelter
  • M Public safety and disaster relief
  • N Recreation and sports
  • O Youth development
  • P Human services
  • Q International affairs
  • R Civil rights and advocacy
  • S Community improvement
  • T Philanthropy and grantmaking
  • U Science and technology
  • V Social science
  • W Public and societal benefit
  • X Religion
  • Y Mutual and membership benefit

Organizations without an NTEE code in the Business Master File form their own unclassified cohort per band.

Minimum size. A cohort needs at least 30 organizations. When a mission group has fewer than that in a band, the adjacent bands are added outward, one on each side at a time, until the minimum is met; the report labels the widened range. If a whole mission group is too thin, the cohort becomes every mission group in that band. In this edition, 157 of 182 cohorts are exact and 25 combine adjacent bands; none needed the mission-wide fallback.

Per-metric counts. Each metric uses the peers that report the figures it needs. Fundraising efficiency and fundraising investment rate count only peers with fundraising expense above zero, so their counts are lower than the cohort size, and the report shows the count under every bar. A metric with fewer than 20 reporting peers shows the value without a percentile.

4. The six metrics

MetricFormulaShown as
Fundraising efficiencyContributions and grants (Part I line 8) divided by total fundraising expenses (Part I line 16b).Dollars raised per dollar spent
Contribution growthContributions in the newest filing compared with two tax years earlier, annualized.Percent change, signed
Contribution share of revenueContributions and grants divided by total revenue (Part I line 12).Percent
Program expense ratioProgram service expenses (Part IX line 25, column B) divided by total functional expenses (column A).Percent
Fundraising investment rateTotal fundraising expenses (Part I line 16b) divided by total expenses (Part I line 18).Percent
Net asset trendNet assets at the end of the newest filing compared with two tax years earlier.Percent change, signed

Growth and trend. Contribution growth compares the newest filing with the filing two tax years earlier and annualizes the change (the square root of the ratio, minus one). Net asset trend is the plain change over the same span. When only two filings are public, the report uses a one-year change and says so. Cohort distributions for these two metrics use only the three-year version.

Bounds. Shares and ratios are clamped to 0 to 100 percent. Fundraising efficiency above $1,000 per dollar is treated as a reporting error and left out of the cohort distribution.

5. How percentiles are placed

For every cohort and metric, the values of the reporting peers are sorted and summarized as 101 quantile points (the 0th through the 100th percentile, linearly interpolated). An organization’s percentile is the share of those points below its value, plus half the share equal to it, rounded to a whole number. Splitting ties this way means a value shared by a third of the cohort lands in the middle of that run rather than at its top.

A percentile is descriptive. The 40th percentile means four in ten peers report a lower figure; it is not a score, a rating, or a judgment about the organization. The report puts the cohort’s 25th percentile, median, and 75th percentile under every bar so the spread is visible.

6. Where the data cannot be compared

Form 990 data is consistent in form and inconsistent in practice. These are the caveats every report carries.

Allocation practices vary

Form 990 lets each organization decide how to split shared costs across program, management, and fundraising. Two organizations doing identical work can report very different fundraising expense. Organizations that report no fundraising expense at all are excluded from the two fundraising ratios, and the report says so.

In-kind gifts count as contributions

Donated goods, food, and services are reported inside contributions and grants. Food banks, medical supply charities, and others with large in-kind programs show very high fundraising efficiency for that reason. The cohort partly corrects for this because peers share the mission group, but read the number with the mission in mind.

Fiscal years differ

A tax year covers fiscal years ending anywhere from January to December of the following year. Peers in one cohort can be up to eleven months apart. The report states the exact fiscal year end for the organization you look up.

The data lags

Filings become public roughly a year to eighteen months after the fiscal year ends, and the newest tax year fills in over several months. This edition uses tax years 2023 and 2024 for the cohorts; an organization whose newest public filing is tax year 2022 is still placed, and the report says which year it used.

Amended and short-period returns

When an organization files more than one return for a tax year, the newest one is used. Returns covering fewer than 350 days are looked up but never placed against peers, because the ratios would not be comparable.

Group returns

A group return consolidates several affiliated organizations. Those filers are looked up but never placed, because there is no comparable single organization.

Negative and tiny bases

Growth and net asset trend need a base year of at least $25,000. Below that, or when the base year is missing, the report shows the figures without a rate rather than a misleading percentage.

7. Refresh and versions

The dataset is rebuilt from the sources above as the newest tax year fills in, at least quarterly. Every report and the aggregate benchmarks page carry the build date and the tax years used, so two reports from different editions can be told apart.

8. What the tool does not do

  • It does not score, rate, or rank organizations, and it publishes no lists.
  • It does not use any data an organization has not already filed publicly.
  • It does not adjust filings, impute missing figures, or estimate anything.
  • It does not name the organizations in a cohort; cohorts are counts and distributions only.

How to cite

StratusLIVE, Nonprofit Peer Benchmark Lookup, edition 2026-09-30, from IRS Form 990 e-file data (tax years 2023 and 2024) and the IRS Exempt Organizations Business Master File. stratuslive.com/tools/peer-benchmark/

Look up your organization.

Headline figures are free and need no email. The full report unlocks with a work email.