The Donor Score Already Inside Your CRM: What Each One Measures

Last updated 2 November 2026. Eight CRMs, read at each vendor's own documentation.

On this page

• Eight CRMs, one blind spot

• What each CRM's score is built from

• What they all have in common

• The part that counts the tag, not the text

• When the score you already have is enough

• When it stops being enough

• A ten minute way to find out which you are

• Where Gratefully fits

• Frequently asked questions

• Sources

Eight CRMs, one blind spot

Most nonprofit CRMs include a donor score. DonorPerfect publishes its formula in full. Bloomerang calls its own proprietary. Virtuous runs eight. Blackbaud sells its scores separately. All of them measure how much, how often, how recently and through which channel. None of these scores takes account of what was actually said.

What each CRM's score is built from

Every figure below was read on the vendor's own documentation on 2 November 2026. Where a vendor does not document something, the table says so rather than filling the gap with a guess.

What each vendor actually names

The detail behind each row, in the vendor's own terms.

DonorPerfect. Recency, Frequency and Monetary value, scored 0 to 100 against your own database and recalculated nightly. The published formula is (0.4 x Recency) + (0.3 x Frequency) + (0.3 x Monetary), and it is made up entirely of those three components.

Bloomerang. Donations, pledges and payments, including in consecutive years. Membership payments. Year on year upgrades. Interactions tagged Channel "In Person" with Purpose "Special Event" or "Volunteer Activity". Interactions flagged "Initiated by Constituent". Website visits. Email opens and clicks in Bloomerang and Mailchimp. Engagement survey responses. Lapsing, year on year downgrades and unsubscribes count against it. Bloomerang also ships a separate Generosity Score, which it describes as the DonorSearch rating for that constituent, built from wealth markers, real estate holdings and philanthropic history.

Keela. The Donor Score uses the average total yearly amount given, the number of years the donor has given, and whether they increased year on year. The separate Engagement Score counts donating, volunteering, attending events and interacting with emails. Note the dates: the Donor Score inputs come from Keela documentation published in 2019 and the Engagement Score entry from 2025, so confirm the Donor Score with Keela before relying on it.

Virtuous. Eight scores including Composite Giving, Lapse Risk, Social, Volunteer and Planned Giving, each 0 to 100 and relative to your own database. First party giving, volunteering, email and web activity and event attendance, combined with third party net worth, income, real estate, demographics and LinkedIn sourced occupation. Engagement itself is measured as Engagement Signals Over Time, Engagement Frequency by Channel, and Last Engagement Date.

Salesforce Nonprofit Cloud. Recency, Frequency and Monetary, calculated by a Data Processing Engine job and available in Enterprise, Unlimited and Developer editions. This is the only score in the table that you schedule and run yourself.

Blackbaud, Raiser's Edge NXT. Annual Gift Likelihood and Major Gift Likelihood, both scored 1 (very unlikely) to 1,000 (very likely). A separate planned giving set covers Bequest, Annuity and CRTL likelihood, for which no score range is published. The knowledgebase article on Annual Gift Likelihood opens with the question "How is this Likelihood calculated?" and the answer describes what the score is for rather than how it is produced.

Four things to know before anyone acts on that table.

Neon and Little Green Light are recorded as "none found", not as "none exists". We searched each vendor's own support site on 2 November 2026 and found nothing. A vendor can ship a feature it has not written up. If you are evaluating either one, ask them directly rather than taking an absence of documentation as an answer.

Blackbaud's scores are a separate purchase. If you are weighing whether the score bundled with your CRM is enough, Raiser's Edge NXT is a different question from the rest of this table, because there is nothing bundled to weigh.

Two Salesforce products get confused here, so be clear which one you are on. The RFM Score above belongs to Nonprofit Cloud. Einstein for Nonprofits is a different package, and Salesforce's own documentation states that the Nonprofit Success Pack app is a prerequisite for it. NPSP and Nonprofit Cloud are separate products with separate roadmaps, so the score available to you depends on which one your org runs.

There is also a score still on some screens that stopped working years ago. Blackbaud's Giving Score, a four band rating released in 2012, was retired in March 2019. Blackbaud's own article notes that organisations can keep the data but no new data has been written since that date, and that users may still be able to open the tool. If you are looking at a Giving Score in an older Raiser's Edge install, you are looking at a number that has not been refreshed since March 2019.

What they all have in common

Three of these scores use giving history and nothing else. DonorPerfect is explicitly Recency, Frequency and Monetary. Keela's Donor Score is amount, years, and change. Salesforce Nonprofit Cloud's framework is RFM by name. For these three, nothing a donor has ever said to you touches the number.

The others widen the inputs without changing their nature. Bloomerang adds email opens, website visits and certain interactions. Virtuous adds channel frequency, last engagement date and a layer of third party wealth and demographic data. Blackbaud adds modelling you cannot inspect.

Every one of them is counting. How much, how often, how recently, through which channel. That is a genuinely useful thing to count, and we have written at length about why RFM still earns its place as a way of ranking a file. It is also the entire range of what these numbers know.

The part that counts the tag, not the text

It is tempting to say that CRM scores ignore your notes. That is not true, and the real answer is more interesting.

Bloomerang does ingest interactions. But its documentation is specific about which ones move the number: interactions carrying the Channel setting "In Person" together with a Purpose of "Special Event" or "Volunteer Activity", or interactions with "Initiated by Constituent" selected. The same article notes that for volunteers it is the number of volunteer activity interactions that counts, not the number of volunteer hours.

Read that twice, because it is the whole point. What moves the score is how the interaction was filed, not what happened in it.

Suppose two notes are logged on the same donor on the same day. Neither is a real record, they are here to make the mechanism visible:

Left voicemail, no answer.

Her husband passed away in March. She asked us to hold off on gala invitations for now, and said she wants to keep giving to the scholarship fund because that is what he cared about.

To the score, those are the same event, worth the same amount, provided both were tagged the same way. If the second one was filed without the right Channel and Purpose, it is worth less than the first.

Be precise about what that does and does not mean. None of the eight scores in the table takes the sentence about the scholarship fund as an input. Several of these platforms separately ship assistants and generative features that can work with written material, and Salesforce Nonprofit Cloud, Virtuous and Keela all have some version of that. Whether a product can read a note is a different question from whether the score is built on one, and it is the score that drives the segment, the list and the ranking your team works from.

This is not a defect. A score is a ranking device and ranking requires countable things. It only becomes a problem when a team starts treating the score as a summary of the relationship, because the part of the relationship that decides what to do next is usually the part that was written in a sentence.

When the score you already have is enough

For a lot of organisations it genuinely is, and anyone telling you otherwise is selling something.

If your main question is "who should get the appeal", the bundled score answers it. Ranking a file by recency, frequency and value is the correct way to decide who receives a mailing, and DonorPerfect's published formula does it as well as anything you could buy.

If one person knows every donor, the score is a convenience rather than a memory. A hundred households held by a single fundraiser who has been there six years does not have an institutional memory problem. It has a capacity problem, and a different number will not fix it.

If you are not yet filing interactions consistently, buying more intelligence is premature. Every product in this table, including ours, is worth less than the discipline of writing things down. If your notes are patchy, that is the thing to fix first, and it costs nothing.

If you are on Salesforce Nonprofit Cloud and nobody has scheduled the job, you may already have a score you are not using. That is worth checking before any other spending.

When it stops being enough

The gap opens in three recognisable places.

When somebody leaves. The score survives a departure. The reason the score is what it is does not. A fundraiser who knew that a donor gives in memory of a specific person, or will not take calls in August, or has a view about the board, takes that with them. We wrote about what that costs in the institutional memory crisis.

When the portfolio is bigger than one person's head. Somewhere past a few hundred relationships the question stops being "who is my best donor" and becomes "which of these three hundred needs attention this month, and why". Ranking answers the first question. It does not answer the second, because the reason lives in the notes.

When a quiet donor is not a lapsed donor. A score built on recency reads silence as decline. The difference between a donor who has drifted and one who asked for space after a bereavement is in a sentence somebody typed, and it is exactly the difference that decides whether your next move is a reactivation appeal or a handwritten note. Our piece on donors at risk of lapsing goes further into the distinction.

A ten minute way to find out which you are

Rather than reasoning about it, test it on your own file. Take your top twenty donors and score each one on four things: when you last had personal contact, how much you actually know about them, whether the next step exists outside somebody's head, and whether the relationship would survive that person leaving.

If you can answer all four for all twenty, your CRM score is doing its job and you do not need anything else. If you cannot, the gap you just found is the one no score in the table above measures.

Our donor relationship health audit is that exercise as a worksheet, with the scoring guide and the thresholds. It is free and it takes about ten minutes.

Where Gratefully fits

Gratefully is not a CRM and does not replace the score in yours. It is an intelligence layer that reads the unstructured side, the notes, documents and history, and produces a ranked list of who needs attention with the reason attached to each name.

It is built for the case in the section above: a portfolio larger than one person's memory, in an organisation where people leave. If that is not your situation, the score you already have is fine.

Frequently asked questions

How does Bloomerang calculate its engagement level?

Bloomerang describes it as a proprietary formula created by Dr. Adrian Sargeant. The company does publish the inputs: donations, pledges and payments, consecutive year giving, membership payments, year on year upgrades, certain tagged interactions, website visits, email opens and clicks, and engagement survey responses. Lapsing, year on year downgrades and unsubscribes count against the score. The weighting between those factors is not published.

Does DonorPerfect's Donor Score use anything other than giving history?

No. DonorPerfect publishes the formula as (0.4 x Recency) + (0.3 x Frequency) + (0.3 x Monetary), producing a score from 0 to 100. It is calculated relative to your own database rather than a national benchmark, and recalculated nightly. The published formula is composed entirely of those three components, with no interaction, email or event component named.

Which nonprofit CRM publishes how its donor score works?

DonorPerfect publishes the complete formula and its weightings. Salesforce Nonprofit Cloud names its framework as RFM and lets the administrator run it. Keela and Virtuous name their inputs without publishing the weighting. Bloomerang names the inputs and describes the formula as proprietary. Blackbaud does not publish the method behind its likelihood scores.

Does Raiser's Edge NXT include a donor score?

Not bundled. Blackbaud's likelihood scores, including Annual Gift Likelihood and Major Gift Likelihood, are delivered through ProspectPoint and Blackbaud Analytics as a separate purchase. A legacy tool called Giving Score was retired in March 2019 and its data has not updated since, although it may still be visible in some installs.

Do CRM donor scores read my notes?

Not the content of them. Bloomerang counts interactions that carry particular Channel and Purpose tags, and Keela's Engagement Score counts email interaction, so a logged note can move a score. What the note says is not a documented input to any of the eight scores reviewed here. The score records that a contact happened, not what was agreed in it. Several of these platforms do also ship AI assistants that can work with written material, which is a separate question from what feeds the score.

Is the donor score in my CRM enough on its own?

For deciding who receives an appeal, usually yes. Ranking a file by recency, frequency and value is the right tool for that job. It stops being enough when the portfolio is larger than one person can hold in memory, when staff turnover means the reasoning behind a relationship leaves with the person, or when you need to tell a quiet donor apart from a lapsing one.

What is the difference between a donor score and wealth screening?

A donor score ranks the people already in your database using their behaviour with you. Wealth screening appends external data about capacity to give. The two overlap: Virtuous combines first party behaviour with third party wealth data in its Composite Giving Score, and Blackbaud's likelihood scores are predictive models sold through the same analytics service as its research products. We compare the screening vendors and what they cost in our guide to nonprofit fundraising software pricing.

Sources

All read 2 November 2026.

• Bloomerang Help Center, "How is engagement level calculated and what does it tell me?", article 12632614
• Bloomerang Help Center, "About the Constituent Summary", article 12632736, for the engagement bands and the Generosity Score
• DonorPerfect, "The Donor Score: What It Is and How to Use It", 11 December 2025
• Keela, donor scoring blog post (published 2019) and CRM glossary entry for Engagement Score (updated July 2025)
• Virtuous Support, "What Do the Virtuous Insights Scores and Values Mean?", article 31792668401933, updated July 2026
• Blackbaud Knowledgebase, "What is Annual Gift Likelihood?", article 43133, last modified 31 December 2025
• Blackbaud Knowledgebase, "What is a Target Analytics Likelihood Score", 14 January 2025, for the requirement to purchase a ProspectPoint service
• Blackbaud Knowledgebase, "What is Major Gift Likelihood?", article 43128, last modified 25 February 2025
• Blackbaud Knowledgebase, "What is a Bequest, Annuity or CRTL Likelihood Score?", article 94948
• Blackbaud Knowledgebase, "What is Giving Score in the Raiser's Edge?", article 53908
• Salesforce Help, "Scoring Frameworks Help Increase Fundraising Success in Nonprofit", and "Einstein for Nonprofits Architecture" for the NPSP prerequisite
• Neon CRM and Little Green Light: vendor support sites searched 2 November 2026, no score documentation located