Running a screening bake-off: the test-file method
Some links on this page are affiliate links. Disclosure
Nothing in wealth screening publishes a rate card. Kindsight, DonorSearch and WealthEngine are all quoted per organization, which means the price you are offered reflects how the negotiation went rather than a market rate.
It also means feature-list comparison is nearly useless. Every vendor claims comprehensive data and accurate scoring. The only comparison that tells you anything is output on people you already know.
Build the test file
Fifty to a hundred constituents whose capacity you can independently verify. Not a random sample — a deliberately constructed one.
Include:
- Board members, whose situations you generally know
- Past major donors at a range of levels
- People whose giving elsewhere is public — named in another organization’s annual report or on a donor wall
- Some you are confident cannot give at a major level, which is the control group most people forget
- A few genuine unknowns, to see what each vendor surfaces
Keep it small enough that you can personally judge every result. A thousand-record file just produces more output you cannot check.
Ask every vendor to screen it
Most will during the sales process. Ask directly, and note that a vendor unwilling to demonstrate output on your own data before a four-figure annual commitment is telling you something.
Give each of them the identical file, with the same fields. Any difference in what you supply becomes a difference in what comes back, and then you are comparing your own inconsistency.
Score them on four things
Did they flag the people you know are capable? False negatives are the expensive failure. A screening tool that misses your obvious prospects will miss the ones you cannot see either.
How many false positives? Names scored high that you know cannot give at that level. A model that flags everyone is not ranking anything, and a gift officer who wastes three visits on bad ratings stops trusting the whole file.
Can they explain a rating? Pick a name and ask how the score was built. This matters more than it sounds. A rating that a gift officer can defend in a portfolio meeting gets acted on; an unexplained number gets quietly ignored, and then you are paying for data nobody uses. Kindsight’s transparency on sourcing is its main differentiator for exactly this reason.
What did they surface that you did not know? The whole point of screening a file rather than hand-researching a shortlist is finding the quiet $80-a-year donor sitting on real capacity. If nothing in the output surprised you, the tool is confirming your existing list rather than extending it.
Expect them to disagree, and use it
Two vendors will rate the same person differently. That is the most informative result in the exercise, and it is not a defect.
The products weight differently on purpose. DonorSearch leads with demonstrated philanthropic history. Kindsight leads with wealth and capacity indicators and shows its sourcing. WealthEngine models wealth across large lists.
Which weighting is right depends entirely on your file. If your donors are modest people who give every year without fail, demonstrated-giving weighting will fit. If they are wealthy people who have not given yet, capacity weighting will. Nobody can tell you which you have except your own data — which is what the test file reveals.
Then negotiate
Tell each vendor you are running the same file past others. Not as a bluff — as a fact.
This is the whole of your leverage in a category with no published pricing, and it works far better before signing than at renewal. Reported figures give you a starting reference: Kindsight from around $4,150/yr for one user and roughly 1,500 screens; DonorSearch from about $4,000/yr for a small file, with per-record screening reported at roughly $1–$2 per record; WealthEngine from around $5,000/yr for a single user plus a one-time implementation fee reported anywhere between $2,000 and $20,000.
Ask about that implementation line specifically. It is the one most likely to appear late.
The option to raise before any of this
If what you actually need is one screen — pre-campaign feasibility, say — you may not need a subscription at all. DonorSearch’s per-record pricing screens a file once with no annual commitment, and a consultant-run screen at roughly $4,000–$5,000 delivers a tiered portfolio with suggested asks rather than a data file.
Our 2026 survey found 95% of shops under $250K raised spend $0 on screening, and half of the $250K–$1M segment still spends nothing. Before running a bake-off, it is worth confirming you are past the point where the category makes sense at all — our wealth screening guide and what a prospect research consultant does both cover that test.
Common questions
Will wealth screening vendors screen a sample file for free?
Most will during the sales process, and you should ask. A vendor unwilling to demonstrate output on your own data before you commit to a four-figure annual contract is telling you something worth hearing.
How large should a test file be?
Fifty to a hundred constituents whose capacity you can independently verify. Large enough that patterns show, small enough that you can judge every result personally. A bigger file just means more results you cannot check.
Should we tell vendors we are testing others?
Yes. Every product in this category is quoted per organization with no public rate card, which means your price reflects the negotiation. A competing bid is the entirety of your leverage, and it works far better before you sign than at renewal.
What if two vendors disagree about the same person?
That is the most useful result in the exercise. Ask each to explain how they arrived at their rating. The one that can explain it is the one your gift officers will trust on the names you cannot verify.
This post is part of the guide to the established fundraising stack ($1M – $5M raised).