Quick answer· The short version
Readers report what checkout showed, by county. The curator reviews every report. A county shows surcharge numbers at 5+ accepted reports and speed numbers at 20+ unobstructed tests, both from the last 6 months. Press reports are a separate, labeled layer. Every cell says "self-reported, not verified by Starlink".
What we collect
| Field | Why we need it |
|---|---|
| State and county | Counties are the smallest area we publish. Smaller areas could point to a household. |
| First 3 ZIP digits (optional) | Helps spot differences inside very large counties. Kept in raw data; never published below county level. |
| What you were doing | New order, reactivation, address change, plan switch or Mini/Roam activation. Surcharges can differ by trigger. |
| What checkout showed and the amount | The point of the record. Whole dollars, $5,000 maximum. |
| Plan type and kit type | The same address can show different results for different plans. |
| Month you saw it | Surcharges change over time. Month, not day, is enough. |
| Speed test (optional) | Download, upload, latency, time of day, tool and obstruction %. Used for speed cells only. |
| Day received | Set by us, used for duplicate checks and review order. No time of day. |
What we never collect
- Street address, precise location or GPS
- Name, email, phone or Starlink account number
- Screenshots (they often show the address)
- Free text of any kind. The form has no comment box
- Your IP address or browser details with a report
- A login or account
Rate limiting without storing your IP: to stop one person flooding a county, the server makes a one-way hash of a secret salt, today's date and your IP address, and counts reports against it for 24 hours. The hash is stored apart from reports, deleted automatically after 24 hours, and can't be turned back into an IP or linked to what you reported.
Nothing goes to SpaceX. Reports are used only for the published county totals and the open dataset. We don't sell or share raw reports.
How reports are reviewed
Form checks
Bots and impossible values are stopped.
Odd amounts wait
Until a second report agrees.
A person reviews
Accept or reject. Never edit.
Counted
Published at 5+ reports.
Each step in detail
Automatic checks at intake
A hidden honeypot field and a minimum fill time catch most bots. Values outside sensible ranges are refused with a message saying why.
Outlier flag
An amount more than 3× or less than ⅓ of a county's current median, or any non-zero amount in a county where every accepted report is $0, is flagged. It waits until a second, independent report agrees, or is rejected after 60 days.
Duplicate check
Identical reports for the same county, trigger, amount and month received the same day are counted once.
Human review
The curator accepts or rejects every report. Reports are never edited. Rejections are counted by reason and published each month.
So far: 0 received, 0 accepted, 0 rejected, 0 pending (as of Oct 5, 2026).
When a county is published
- Surcharge cell
- n ≥ 5 accepted reports in the last 6 months. Shows median, range, most common trigger and dates.
- Speed cell
- n ≥ 20 tests in the last 6 months, excluding tests with more than 1% obstruction. Shows p25, median and p75.
- State page
- Indexed by search engines only when at least 3 of its counties are published.
- Below threshold
- "Not enough reports yet" and the count so far. Never an estimate, never a neighbor's numbers.
Map colors use five steps: $0, $1–249, $250–499, $500–999, $1,000+. The scale is readable with common color-vision differences. "Not enough reports" is hatching, never a color, so it can't be mistaken for a value.
The press layer
News articles sometimes quote a surcharge for a city or state. We log each one with its outlet, link, publication date and the quoted amount, as a separate layer. Press amounts show as "unverified" until a matching crowd report or a Starlink support article confirms them, and they never enter a county's crowd median.
Limits of this data
- Self-selection. People who saw a surprising surcharge are more likely to report than people who saw $0. We ask for $0 results specifically to balance that, and we say so.
- Lag. A rolling 6-month window smooths noise but responds slowly when a surcharge changes quickly.
- Address-level variation. Two addresses in one county can see different results. A county median is a guide, not your quote.
- No verification by Starlink. We can't check a report against Starlink's systems, and we don't try to.
Method changelog
- · method-1
Method published. Surcharge cells publish at n ≥ 5 reports and speed cells at n ≥ 20 tests, in a rolling 6-month window. Speed tests with more than 1% app-reported obstruction are excluded. Press reports are a separate layer and never enter crowd medians.