Methodology

IRS Form 990 is the annual information return that tax-exempt organizations file. It's a public record. It includes revenue, expenses, assets, officer names and compensation, program descriptions, and governance information. Think of it as the financial X-ray of a nonprofit — not a full diagnosis, but a clear picture of the skeleton.

Harbor Commons ingests data from three IRS-provided sources:

• –EO Master File — the IRS Exempt Organizations Business Master File, used to build the universe of sailing-adjacent nonprofits.
• –TEOS Index — the IRS Tax Exempt Organization Search annual filing index, used to identify available e-file returns by year.
• –E-file XML — structured machine-readable 990 filings. These are accessed via the GivingTuesday 990 Data Lake, an open mirror of IRS TEOS e-file submissions.

All financial figures come directly from the filing XML. Harbor Commons does not introduce external benchmarks, sector averages, or estimated values.

The IRS typically publishes e-file XML 6 to 18 months after the filing date, which itself may be 6 months or more after the organization’s fiscal year end. In practice:

• –A club with a December 31, 2024 fiscal year end may not file until September 2025.
• –That filing may not appear in TEOS until mid-2026.
• –Harbor Commons reflects data as of its most recent index refresh.

Unknown is a valid output. Where data is absent, we say so — we don’t interpolate.

A treasurer or commodore reviewing Harbor Commons may be looking at a real filing that is still a year or more behind the organization’s current day-to-day reality. Harbor Commons is best used for historical pattern recognition, not as a live dashboard.

Peer comparison is the most powerful — and most easily misused — feature of any 990 tool. A few principles:

• –Align to the same tax year before discussing differences. Revenue and expense figures are not comparable across different filing years.
• –Compare totals first, then read narratives. Part VIII line items and program descriptions require context from inside the organization.
• –You’re looking for structural peers, not regional norms. An org of similar size, type, and program mix anywhere in the country may be more comparable than a nearby club with a different model.
• –Form type matters. A 990 and a 990-EZ disclose different depths of detail. Treat mixed-form comparisons as approximate.
• –Signals flag patterns — they don’t supply explanations. The filing shows a shift; people inside the organization supply the meaning.
• –Labor totals include employer benefits and payroll taxes. Per-person disclosures (Part VII-A) are already inside the Part IX total — they are not additive and aren't directly comparable to the whole-org figure.
• –Professional and consulting fees (Part IX lines 11a–11g) cover payments to outside firms and independent contractors — these are reported separately from the W-2 labor total and are not included in it.

See Filing Signals for what each signal means — and what it does not conclude.

We welcome context and corrections about how a signal should be interpreted, whether a data field is being displayed incorrectly, or whether something about an organization’s situation is missing from its page.

We do not alter IRS source filings. We correct our parsing, labeling, and presentation when we’re wrong. Send us a signal →

See Updates for a log of recent data additions, method notes, and confirmed gaps.