LOG-005 ·
The Sovereignty Score: One Number I Compute Every Quarter
I built a metric for how much of my life runs on infrastructure I control, and I have been tracking it for nine quarters.
- Words
- 649
- Est. read
- 2.8 min
- Confidence
- 0.84
- Topics
- measurement, sovereignty, metrics
What gets measured gets owned, but ownership itself also needs a measurement, or the project of decentralizing your life becomes a mood instead of a program. So in 2024 I built one: the Sovereignty Score. I have computed it every quarter since, nine data points so far, and today I am publishing the method so anyone can compute their own and so my future numbers can be checked against a stated formula rather than a shifting one.
The method
List every system the household depends on. Mine has 23, from money movement to photo storage to the garage door. For each system, assign a control rating:
- 2 points: sovereign. I hold the asset and can operate it if every vendor involved disappears. Example: photos, which exist in three copies on media I own.
- 1 point: portable. A vendor operates it, but I can move to a competitor within one weekend without losing data or identity. Example: email on my own domain. The provider is a replaceable part.
- 0 points: captive. The vendor holds the asset, the identity, or both, and leaving means losing something. Example, to my continuing irritation: prescription records.
Each system also gets a weight from 1 to 3 for how much damage its loss would do. Money movement is a 3. The robot vacuum's app is a 1. The score is the weighted sum of ratings divided by the weighted maximum, expressed as a percentage.
Two design decisions worth defending. First, weighting matters because an unweighted score rewards decentralizing trivia. Second, "portable" earns half credit deliberately: exit rights are worth something real even when you do not exercise them, a principle economists have written about for fifty years under the name of contestable markets. The credible threat of leaving disciplines the vendor.
Nine quarters of data
| Quarter | Score | Notable change |
|---|---|---|
| 2024 Q2 | 31.2% | Baseline. Grim. |
| 2024 Q3 | 34.8% | Email to own domain |
| 2024 Q4 | 41.5% | 3-2-1 photo backups |
| 2025 Q1 | 44.1% | Second bank, self-custody started |
| 2025 Q2 | 49.6% | NAS replaces cloud storage |
| 2025 Q3 | 52.3% | 5G failover, local DNS |
| 2025 Q4 | 55.0% | Notes to local-first app |
| 2026 Q1 | 58.7% | Password manager self-hosted |
| 2026 Q2 | 61.4% | Solar battery commissioned |
From 31.2% to 61.4% in nine quarters, a mean improvement of 3.4 points per quarter, though the trend is visibly decelerating. The early points were cheap: any weekend project moved the number. What remains is the expensive tail. Banking rails, government identity, prescriptions, the property title. Some of those may never leave zero, and the score is honest about that in a way my optimism is not.
What the metric taught me that intuition did not
First: I systematically overestimated my own sovereignty before scoring it. Pre-baseline, I would have guessed 50%. The audit said 31.2%. Feelings of independence and measured independence are different quantities, and the gap between them is exactly where complacency lives.
Second: the biggest single-quarter jumps came from boring moves. Backups and email outscored anything involving a blockchain, by a wide margin, at a fraction of the discourse.
Third: a plateau is information. When the score stalls two quarters running, the remaining work is either genuinely hard or genuinely not worth it, and the weights tell me which.
The caveats, stated
This is a self-graded exam and the rubric has judgment calls. Is a service with an export button portable if I have never tested the export? (My rule: untested exports score 0. Test your exports.) The weights encode my household's values, not yours. And a score of 100 is neither achievable nor desirable; I am not interested in smelting my own silverware.
I hold the target of 70% by the end of 2027 at 0.84 confidence. When 2027 Q4 arrives, the number will be here in the log, next to this entry, and one of us (the forecast or me) will be embarrassed. That is the arrangement I prefer with all my beliefs.