How the Protection Score Is Calculated
The categories, weights, and limits of SpoofKiller's 0 to 100 domain score.
Updated
The Protection Score is a number from 0 to 100 that summarizes how well a domain’s email authentication is configured, enforced, and watched. It is built from 7 categories worth 100 points between them. Every point is traceable to a DNS record or a figure from your DMARC reports, and this page lists all of them.
What the score is not
The 7 categories
| Category | Weight | Question it answers | Measured from |
|---|---|---|---|
| DMARC configuration | 20 | Is there one valid DMARC record, with reporting? | DNS scan |
| DMARC enforcement | 20 | How strong is the published policy? | DNS scan |
| SPF configuration | 15 | Is the SPF record valid and within its limits? | DNS scan |
| DKIM authentication | 15 | Is mail carrying a DKIM signature aligned with the domain? | Reports, or DNS scan when there are none |
| Authentication performance | 15 | How much reported mail actually passes DMARC? | Reports only |
| DNS security | 10 | Are DNSSEC, MX, and transport security in order? | DNS scan |
| Monitoring coverage | 5 | Is the domain verified, reporting, and receiving reports? | Account state and reports |
| Total | 100 |
Each category produces a ratio between 0 and 1, and earns that fraction of its weight. The sections below give the ratio for each. “Reports” means DMARC aggregate reports for the last 30 days; they count only for verified domains.
DMARC configuration (20 points)
A single well-formed record with a reporting address earns the full weight. Problems subtract from it:
| Condition | Effect on the ratio |
|---|---|
| No DMARC record published | 0 |
| More than one DMARC record, or a record with no policy | 0 |
v=DMARC1 is not the first tag | −0.50 |
No aggregate reporting address (rua) | −0.30 |
| An external report destination has not authorized the domain | −0.15 |
| A reporting URI is malformed | −0.15 |
| Invalid or duplicated tags | −0.15 |
Deductions add up and the result is held between 0 and 1. The DMARC checker reports the same conditions for any domain.
DMARC enforcement (20 points)
This category rewards the policy receivers are actually asked to apply, including the pct tag, which limits the share of failing mail the policy covers.
| Published policy | Ratio | At pct=100 | At pct=25 |
|---|---|---|---|
p=reject | 0.50 + 0.50 × pct | 1.00 | 0.625 |
p=quarantine | 0.35 + 0.35 × pct | 0.70 | 0.4375 |
p=none | 0.15 | 0.15 | 0.15 |
| No policy | 0 | 0 | 0 |
If the domain enforces a policy but sets sp=none, leaving subdomains unprotected, 0.20 is subtracted. A domain at p=none is not being penalized for caution: monitoring first is the right order of work, as None vs Quarantine vs Reject explains. The category simply reports that enforcement has not started.
SPF configuration (15 points)
The SPF check produces its own health score from 0 to 100 by walking the record and every include, counting DNS lookups against the limit of ten, and looking for errors such as multiple records, syntax errors, void lookups, and +all. The ratio is that score divided by 100. No SPF record earns 0. The SPF checker shows the same analysis, and Common SPF Errors covers the fixes.
DKIM authentication (15 points)
DKIM is measured from what mail actually does when that is known, and from published keys when it is not:
- With reports: the ratio is the share of reported messages that carried a DKIM signature aligned with the domain. If 96% of messages were DKIM-aligned, the ratio is 0.96.
- Without reports, keys found: the ratio is the DKIM check’s health score for the keys found, divided by 100. That score reflects problems such as weak, malformed, or revoked keys.
- Without reports, no keys found: the category is not measured. DKIM selectors cannot be enumerated from DNS, so “no keys at the selectors checked” does not mean “no DKIM.” Add the selectors your sending services use and the category becomes measurable.
Authentication performance (15 points)
This is the DMARC pass rate across reported messages, on a deliberately steep scale: a pass rate of 90% or lower earns nothing, 100% earns everything, and the points rise in a straight line between. A 95% pass rate has a ratio of 0.5; 99% has 0.9. The steepness is intentional. On a domain sending a hundred thousand messages a month, 95% means five thousand messages that would be quarantined or rejected under enforcement. Without report data the category is not measured.
DNS security (10 points)
| Component | Adds to the ratio |
|---|---|
| DNSSEC is signed and validating | 0.40 |
| MX records exist with no high-severity problems | 0.30 |
| MTA-STS policy published, in enforce mode | 0.15 |
| MTA-STS policy published, in another mode | 0.10 |
| TLS-RPT record published | 0.10 |
| At least one CAA record | 0.05 |
The DNS security checker reports each of these for any domain.
Monitoring coverage (5 points)
| Component | Adds to the ratio |
|---|---|
| Domain ownership is verified | 0.30 |
| The SpoofKiller reporting address is in the DMARC record | 0.30 |
| A report was processed in the last four days | 0.40 |
A configuration nobody is watching decays, so a small part of the score is reserved for watching it. In the free public tools, where there is no account, this category is not measured.
Unmeasured categories and scaling
A category is left unmeasured, rather than scored as zero, when there is not enough data to judge it: no reports yet, no DKIM keys at the selectors checked, or a DNS lookup that failed. Scoring those as zero would punish a new domain for being new and a domain for a resolver’s bad afternoon.
The score is then scaled over the categories that were measured:
score = round( points earned ÷ total weight of measured categories × 100 )A score calculated this way is marked as partial in the dashboard and the API. An example, for a newly verified domain with no reports yet:
| Category | Weight | Ratio | Points | Why |
|---|---|---|---|---|
| DMARC configuration | 20 | 1.00 | 20 | Valid record with a rua address |
| DMARC enforcement | 20 | 0.70 | 14 | p=quarantine at pct=100 |
| SPF configuration | 15 | 0.90 | 13.5 | SPF check scored 90 of 100 |
| DKIM authentication | 15 | not measured | — | No keys at the selectors checked, no reports |
| Authentication performance | 15 | not measured | — | No reports yet |
| DNS security | 10 | 0.40 | 4 | MX healthy, TLS-RPT published; no DNSSEC, no MTA-STS |
| Monitoring coverage | 5 | 0.60 | 3 | Verified, reporting address confirmed, no report yet |
| Measured | 70 | 54.5 | 54.5 ÷ 70 × 100 = 78 |
When reports arrive, the two unmeasured categories join the calculation and the score moves, up or down, to reflect what the mail is really doing. Expect that movement in the first week.
Status bands
| Score | Status | Summary shown |
|---|---|---|
| 90 to 100 | Strong | Email authentication is well configured and enforced. |
| 75 to 89 | Good | The fundamentals are in place; a few improvements remain. |
| 50 to 74 | Needs Improvement | Important protections are missing or not enforced. |
| 0 to 49 | High Risk | The domain is poorly protected against impersonation. |
Recommended fixes
Each category carries its own list of improvements. The dashboard shows up to five, ordered by how many points the category is leaving unearned, so the first item is the one worth the most. A fix can raise the score without making enforcement safe, and the reverse; whether to tighten the policy is a separate question, answered by enforcement readiness.
Limits of the score
- Report-based categories reflect mail seen by receivers that send aggregate reports, which is not all mail.
- DKIM without reports reflects only the selectors checked.
- The score describes one domain. Subdomains added separately are scored separately.
- The weights are SpoofKiller’s judgment and may be tuned. If they change, this page changes with them.
