What every profile carries

Every finding in a DonorGraph profile is a claim with a document behind it: a property with its deed and parcel record, a board seat with the Form 990 line that names the officer, a securities holding with the SEC filing, a contribution with the FEC or state campaign-finance line. When a gift officer is asked where a number came from, the answer is to open the record. The profile is not a rating to be trusted; it is evidence to be checked.

What we commit to

CommitmentWhat it means in the productWhat it gives a researcher
Cited evidence per fieldEvery asserted value links to its primary record: deed, Form 990 line, SEC filing, FEC or state campaign-finance line.A finding that can be audited on the spot, by anyone in the room.
Identity resolution against public registriesMatches are made against IRS BMF/990, FEC, deed and SEC records using name, place, household, self-reported employer and organizational ties; the matching record is shown.Protection from same-name collisions, wrong-person property and mis-attributed board seats.
Evidence tiers on every findingEach value is labeled confirmed, possible or none, in the same vocabulary the profile uses on screen.A profile that can be read in a minute, with the strong findings and the open questions already separated.
One-shot, held-out evaluation with negative controlsA ground-truth set the system never trains or tunes on, scored in a single pass with no retries, seeded with fabricated identities a correct system must reject.Confidence that the method is measured against the world rather than against itself.
Refusal when evidence is thinFields read no evidence found with the reason, rather than inferring a value from ambiguous records.A blank you can act on, instead of a number you have to disprove.
One-click correctionAny field can be corrected; the correction is stored for your organization and applied to future screenings.An error caught once stays fixed, for everyone on the team.

Where the data comes from

DonorGraph is built on primary public records rather than licensed consumer profiles: IRS Business Master File and Form 990 filings, SEC EDGAR filings, federal FEC and state campaign-finance filings, county deed and parcel records, and the organizational registries that connect people to boards and businesses. Each has a known structure, publisher and refresh cycle, which is what makes citation possible: a fact can be linked to a document a researcher can open.

How identity resolution works

The hard problem in screening is not finding records; it is deciding which records belong to the person in front of you. DonorGraph resolves identity in layers. Structured anchors come first: name and location, employer as self-reported on a campaign-finance filing, a household member listed as a deed co-owner, an organization the person is known to serve. Corroboration across sources follows: a Form 990 board seat carries more weight when that organization also appears in the person's giving history. Where the records cannot separate two plausible candidates, the profile presents the ambiguity instead of picking one. Every accepted match shows its supporting record, and a researcher can reject it in one click.

Evidence tiers

Every finding is labeled with the strength of the record behind it, in the vocabulary the product uses on screen.

The tiers are what let a profile be read quickly: work the confirmed findings, spend judgement on the possible ones, and know exactly where the public record runs out.

How we test

Held-out ground truth

We maintain a set of donor records with independently established facts. The system does not train on it, and matching and prompting are never tuned against it. The set is run once, in one shot, with no retries and no manual intervention, and the output is compared field by field to the ground truth.

Synthetic negative controls

The evaluation also includes identities that do not exist and real names paired with the wrong location or employer. A system that returns a capacity number for a person who does not exist is measuring its own confidence. The negative controls are how we know refusal is working.

Field-level review

Property, board service, securities, political giving and identity match are reviewed separately, because they fail differently and are fixed differently. Where a field is weak the default is to refuse rather than guess. That is a design rule, not a temporary setting.

What we deliberately do not do

Correction is part of the method

Every field has a one-click correction. It is stored for your organization, applied when that donor is screened again, and fed back as a signal to identity resolution on similar cases. Screening errors cluster around common names, shared family property and similarly named organizations, so a correction loop turns each caught error into a rule.

Four questions to ask any screening vendor

Can I see the source document behind this value? What does the system return for a person who does not exist? What happens when two people share a name in the same city? And when I find an error, what changes?

DonorGraph's answers are on this page and in the product. Pricing is on the pricing page; the comparisons with DonorSearch, iWave and Blackbaud WealthPoint set out the differences fact for fact. The fastest check is the one you run yourself on the trial page.

Frequently asked questions

What does 'cited evidence per field' mean?

Every value in a DonorGraph profile links to the primary record it came from: the deed, the Form 990 line, the SEC filing, the FEC or state campaign-finance line. A gift officer asked where a number came from opens the document.

How does DonorGraph decide that a record belongs to my donor?

It runs against structured public registries: the IRS Business Master File and Form 990 officer lists, FEC filer records, county deed and parcel records, and SEC filings. It anchors on name, location, deed co-owners, self-reported employer and organizational ties, then corroborates across sources. Every accepted match carries the record that supports it.

What are the evidence tiers?

Confirmed, possible and none. Confirmed means a primary record names the person and the identifiers line up. Possible means the identifiers do not close the match, and the profile states the ambiguity and links the record. None means the records do not support a conclusion, and the field reads no evidence found with the reason.

How is the system tested?

Against a held-out set of donors with independently established facts that the system never trains or tunes on, run once in a single pass with no retries, and seeded with synthetic negative controls: fabricated and mismatched identities that a correct system must return nothing for. Results are reviewed field by field.

What does DonorGraph deliberately not do?

It does not present modeled net worth as fact, and it does not match on name alone. A wealth estimate produced by a model is a statement about the model; DonorGraph reports the assets it can document. A name without a corroborating identifier is a lead, not a finding.

What happens when I correct a field?

It is stored for your organization, applied to that donor's future screenings, and used as a signal for identity resolution on similar cases. It does not change what other organizations see.

See a complete screening on a donor you already know.

Try 3 screenings free → See pricing
Sources and notes
  1. Primary record sources described above are public: IRS Business Master File and Form 990 (irs.gov), SEC EDGAR (sec.gov), FEC bulk data (fec.gov), state campaign-finance portals, county recorder and assessor records.
  2. This page describes method. Evidence tiers, refusal behaviour and one-click correction are visible in the product on any screening, including the three in the free trial.