Knowledge Base
Digital Karma Score: What the Seven Signals Measure
The Digital Karma Score turns seven site checks into one number from 0.00 to 1.00. More importantly, it shows which signals were measured, which were declared, and which still need evidence.
The Seven Digital Karma Score Signals
- Schema Coverage (20%) ... how completely the site explains page meaning with Schema.org.
- Content Freshness (15%) ... whether the public discovery layer still reflects the current site.
- AI Endpoints (25%) ... completeness and validity of the required public files.
- Federation Presence (15%) ... quality of peer relationships and network context.
- External Links (10%) ... relevance and quality of wider web references.
- Technical Quality (10%) ... HTTPS, mobile behavior, and general implementation health.
- Dataset Quality (5%) ... usefulness and completeness of published structured assets.
Worked Example: Seven Signals Become a 0.82 Score
Suppose Example Site has strong endpoints and schema but still needs better datasets and outside references. Each signal is multiplied by its weight.
| Signal | Signal value | Weight | Score contribution |
|---|---|---|---|
| Schema Coverage | 0.90 | 20% | 0.180 |
| Content Freshness | 0.80 | 15% | 0.120 |
| AI Endpoints | 1.00 | 25% | 0.250 |
| Federation Presence | 0.70 | 15% | 0.105 |
| External Links | 0.50 | 10% | 0.050 |
| Technical Quality | 0.90 | 10% | 0.090 |
| Dataset Quality | 0.50 | 5% | 0.025 |
| Total | 100% | 0.820 |
The score is 0.82 because the weighted contributions add up to 0.820. Improving a heavily weighted signal such as AI Endpoints changes the total more than the same improvement to Dataset Quality.
Measured, Declared, and Unknown Are Different States
| Status | What it means | How to read it |
|---|---|---|
| Measured | A current check produced the value from public evidence. | The score can name the check and when it ran. |
| Declared | The site or its generator published the value without an independent current measurement. | Use it as a stated value, not as outside proof. |
| Unknown | The evidence is missing, stale, or unavailable. | Keep it unknown until a real check can replace it. |
The current specification records an unresolved limitation: several federation sites still publish static defaults for some signals. A transparent score keeps those declared values visible instead of dressing them up as measurements.
How Digital Karma Badges Use the Score
- Karma Certified requires all seven public artifacts to validate, including the rendered-schema receipt, plus a score of 0.70 or higher.
- Karma Pro starts at 0.85 and also requires useful datasets plus bidirectional peer links.
- Karma Elite starts at 0.95 and also requires automation, diagnostics, and entity files.
A threshold alone does not earn the badge. The site must meet the added requirements listed beside that threshold.
What a Digital Karma Score Can and Cannot Prove
The score can show whether the public implementation is complete, current, connected, and backed by useful structured information. It can also show which part of the build is holding the total down.
It cannot prove that an AI system used the data, cited the site, or sent a human visitor. Those outcomes need their own crawler, referral, citation, and search evidence.
How Site Owners Raise the Score
Start with the gaps that have public receipts. Complete the required files, keep the sitemap and catalog aligned, render the catalog across every canonical page, and publish a current zero-failure receipt.
Then improve the signals that need deeper work. Add useful datasets, maintain real peer relationships, strengthen technical quality, and build references that exist beyond the site itself.