How to read arts finance reporting and audience claims

What to take away

This arts finance reporting method lets a reporter examine a nonprofit arts organization without turning one filing into a verdict. Financial condition, program reach and artistic quality overlap, but they are not interchangeable.

Photo and credit

The photograph shows one exterior entrance on one date. It does not establish the institution's current hours, finances, admission policy or complete accessible route.

Step 1: match entity and period

Write down the legal name and the employer identification number where public. Also write down the fiscal-year end and the public brand. Note any subsidiary or related foundation and any fiscal sponsor. Finally, write down the venue owner. Then check whether consolidated statements include entities missing from the tax return.

Create columns for document issue date and period covered. A report published in 2027 may describe a fiscal year ending in 2026. Comparing it with a calendar-year program total needs an adjustment, not a footnote.

Step 2: read the filing structure

On Form 990, start with these areas:

Then follow the relevant attached or referenced schedules, rather than stopping at the first-page totals:

For example, if Part VIII reports contribution revenue, do not treat that entry alone as proof that a particular grant was awarded, paid in cash, unrestricted or available to spend. Reconcile it with the grant agreement, payment records, restrictions and audit notes; mark details you cannot verify as unknown.

The IRS instructions state that Form 990 and Form 990-EZ carry the information required from covered tax-exempt organizations and are generally open to public inspection. They also separate governing documents, conflict policies, financial statements and disclosure rules. Use them for field meaning, then read the organization's own filing.

For a state-level cross-check, search the relevant state attorney general's charity-registration records, or the state's designated charity regulator. Registration and filing requirements vary by state, so identify the jurisdiction and period rather than treating a state record as a second copy of Form 990.

Do not infer misconduct from a high or shifting line before you understand classification, one-time transactions, donated services, capital activity and accounting method.

Step 3: compare the audit

Read the auditor's opinion, then the statements, cash flows, functional expenses and notes.

Consider these areas when reviewing the financial statements:

Reconcile differences by consolidation, period, accounting treatment and classification. Record management's explanation, and bring in independent accounting expertise when the conclusion turns on technical interpretation.

Step 4: trace a grant

Build a grant ledger. One row per award, with the fields below filled or marked unknown.

Field	What it settles

| Grantor and legal recipient | Establishes who provided and received the award, including any fiscal sponsor. | | Award date and amount | Separates the award from a multi-year ceiling or later payment. | | Cash received and date | Shows what was paid and when. | | Revenue recognized and fiscal year | Shows when the amount was recognized in the accounts. | | Project period | Identifies the period the award covers. | | Restrictions and conditions | Shows what the recipient may use the funds for and whether conditions remain. | | Source document and page | Lets a reader verify the entry. |

A press release usually establishes two or three of these. Separate a multi-year ceiling from one year's payment, and separate a grant to a fiscal sponsor from money the project controls directly.

Step 5: define audience units

Ask what the numerator counts, and write the answer down in the institution's own words. Tickets issued, ticket scans, unique people, and visits are four different numbers. Program completers, estimated passersby, and online sessions are three more numbers.

Ask about duplicates, complimentary access, no-shows, staff and artist attendance, canceled programs, member visits and multi-venue totals. Preserve the raw category names before combining anything.

Step 6: audit the survey claim

Obtain the questionnaire, the recruitment method, the field dates, the sample size and the completion rule. Get the response rate, the weighting, the geography, the comparison year and the uncertainty. Check whether wording or delivery changed between years.

Pew Research Center's description of its U.S. survey methodology shows what full disclosure looks like: a probability-based panel, a stated target population, and a published topline questionnaire with exact wording and response options. Expect a venue's exit survey to fall short, then establish how far.

Turn "visitors felt welcome" into a bounded statement: "among 312 respondents to a voluntary exit survey run over four weekends, 81 percent chose 'very welcome.'" Then say who the method leaves out.

Step 7: interview with a reconciliation table

Send the institution a table of apparent differences. Ask which entity, period, unit and method each number uses. Request the supporting schedules or a correction, and set a clear response deadline.

Step 8: write the claim ledger

For every figure, store its source, page, definition, period and unit. Also store its calculation, contact response and confidence level. Report why two numbers differ instead of defaulting to the larger or the newer one.

Common questions

Does negative net income mean an institution ran out of cash?

Not necessarily. Read the cash flow statement, liquidity, restrictions, debt and the notes before drawing that conclusion. A deficit year and a cash crisis are separate findings, and only one of them belongs in a headline.

Is an award announcement revenue?

Not automatically. The terms, conditions, project period, accounting treatment and payment status all decide that. A pledge announced in one fiscal year may be recognized in another, or not at all if it carries a condition the organization has not met.

Can visits be called visitors?

Only if unique people were measured. Otherwise use visits or admissions, and say which one the source counted. Mixing the two across years produces a trend that does not exist.

What if the institution will not share survey questions?

Report the published result with that missing-method limit stated in the sentence, or omit the claim if it cannot be assessed responsibly. The verification checklist is the fallback when a source goes quiet.