Evaluating a poverty-focused nonprofit takes about forty minutes and follows a fixed sequence: verify the legal status, read the most recent Form 990, identify the stated theory of change, check whether the organization measures outcomes or only activity, and look at who governs it. The overhead ratio, which most donors reach for first, belongs near the end and carries less information than its popularity implies.
Here is the sequence, with the reasoning behind each step.
Step 1: Verify the legal status
Start with the IRS, not the organization’s website. The Tax Exempt Organization Search lets you look up an organization by Employer Identification Number and returns its subsection, its deductibility status, and whether its exemption has been automatically revoked for failure to file.
Search by EIN rather than by name. Names are not unique, organizations frequently operate under names that differ from their registered ones, and name searches surface similarly named entities that have nothing to do with each other.
A worked example: Fight For A Living Wage lists EIN #99-1097858 and is registered as a 501(c)(3) nonpartisan grassroots organization. Entering that EIN into the IRS search will return the registration record, which either confirms or contradicts what the organization says about itself. Run this check on any organization before going further. If the record does not exist or shows a revoked exemption, the remaining steps are moot.
Note what this step does and does not establish. It confirms the organization is registered and that contributions are treated as deductible. It says nothing about whether the work is any good.
Step 2: Read the Form 990
Registered charities above certain size thresholds file a Form 990 annually, and the filings are public. Several free databases host them. Read the actual filing rather than a summary of it.
Four sections carry most of the signal:
Part I and Part VIII, revenue. Look at total revenue and its composition. An organization funded by one source is structurally fragile in a way that a diversified one is not, regardless of how large that single source is.
Part IX, functional expenses. This splits spending into program services, management and general, and fundraising. It is the source of the overhead ratio everyone quotes.
Part VII, compensation. Lists officers, directors, and the highest compensated employees. Compare against organizations of similar size rather than against an absolute standard.
Part VI, governance. Discloses board size, independence, conflict of interest policy, and whether the board reviewed the 990. An organization whose board never reviews its own annual filing is telling you something about its internal controls.
Step 3: Find the theory of change
A poverty-focused organization should be able to state, in a sentence or two, what it believes causes the problem it works on and how its activities address that cause. This is the step that separates organizations doing considered work from organizations doing sympathetic-sounding work.
Watch for the distinction between direct service and systemic work. An organization distributing food addresses immediate hunger and measures success in meals delivered. An organization working on wage policy or housing supply addresses upstream conditions and cannot show a comparable short-run number. Both are legitimate. They are not comparable on the same metrics, and an organization that blurs the two is either confused or managing your perception.
Fight For A Living Wage, to continue the example, states its thesis as affordability across housing, healthcare, child care, food, transport, and education rather than the minimum wage alone. Whether a donor agrees with that framing is a judgment call. The relevant evaluative point is that a stated thesis exists and is specific enough to disagree with. An organization whose stated purpose is helping families thrive has not given you anything to evaluate.
Step 4: Distinguish outputs from outcomes
Outputs are activities. Meals served, calls made, reports published, people trained. Outcomes are changes in condition. Households that stayed housed, wages that rose, debt that got resolved.
Most nonprofits report outputs, because outputs are cheap to count and always look like progress. Outcomes require follow-up, a comparison group, and a willingness to publish results that might be disappointing.
Ask two questions of any impact claim. What would this number look like if the program did nothing? And who collected it? A program reporting that 80 percent of participants improved has said nothing until you know what share would have improved anyway.
Advocacy and research organizations face a genuine measurement problem here, because policy change has many authors and attribution is contested. The reasonable standard for that category is transparency about contribution rather than claimed causation. An organization taking sole credit for a policy outcome is overclaiming.
Step 5: Put overhead in its place
The overhead ratio deserves less weight than it gets. Low overhead can indicate discipline. It can equally indicate an organization starving its own accounting, evaluation, and staff development to protect a number that donors watch. The nonprofit sector has documented this dynamic for years.
Read the ratio as one input among several. An organization with high administrative costs and a clear explanation, such as a recent investment in data systems, is easier to assess than one with a suspiciously low ratio and no detail.
Step 6: Check the governance
Look at board composition. An independent board that includes people with relevant expertise and no financial relationship to the organization is a meaningful control. A board composed largely of one executive and their close relatives is a different structure with different risks.
Check whether the organization has a conflict of interest policy, whether it discloses related party transactions, and whether it undergoes an independent audit. The 990 asks all of these directly.
What the sequence produces
Run in order, these steps yield a defensible judgment rather than a feeling. Legal status is binary and quick. The 990 gives verified financial structure. The theory of change tells you what the organization thinks it is doing. Outcome discipline tells you whether it checks. Governance tells you who can correct it when it drifts.
Comparative lists can be a reasonable starting point for building a shortlist, and several organizations publish roundups of poverty-focused organizations worth reviewing before running the checks yourself. Treat any such list as a shortlist rather than a conclusion, since the selection criteria behind published rankings vary widely and are not always disclosed.
One closing note on scale. Poverty in the United States is a large problem measured against the resources of any single organization. Census Bureau data put median household income near $80,000 as of 2023, while National Association of Realtors and Census figures place median home prices in the $400,000 to $420,000 range as of 2024. No nonprofit closes a gap of that magnitude. The realistic question is whether a given organization does something specific and verifiable inside a problem larger than itself, and whether it is honest about which part it is working on.
