1. My county says "not enough reports yet"
Quick answer
It means fewer than 5 accepted surcharge reports (or 20 speed tests) from your county in the last 6 months. That's a deliberate floor, not a missing page.
With two or three reports, one unusual quote swings the median wildly. Five is the smallest number where a median starts meaning "typical here" rather than "one person's day". For speeds, results vary far more from test to test, so the floor is 20.
What you can do: check your own checkout (it's the only real quote), and add what you saw. The lookup shows how many reports your county has so far.
2. A number looks wrong, or fake
Some reports are typos (a hypothetical example: $2,500 typed instead of $250), some are honest but unusual, and a few are deliberate. Here's what happens to each:
- Out-of-range values are refused at the form with a message, before they reach us.
- Unusual values (more than 3× or less than ⅓ of the county median, or any surcharge in an all-$0 county) are held until a second independent report agrees.
- Duplicates sent on the same day are counted once.
- Floods from one connection are rate-limited for 24 hours, without storing the IP address.
A median also resists outliers: in a hypothetical county with five $250 reports, one $1,500 report doesn't move it. If you still think a published number is wrong, tell us. Every correction is logged on the corrections page.
3. A press number and a crowd number disagree
Quick answer
They measure different things. A press number is one dated quote, usually for a city or metro. A crowd number is the median of many reviewed reports for one county. Both can be accurate and still differ.
| Source | What it says | How to read it |
|---|---|---|
| Press | Hypothetical: "Buyers in Example City saw $500" (article from 14 months ago) | Someone in the metro saw that amount on that date. It may be higher or lower now. |
| Crowd | Hypothetical: Example County, median $250, n = 9, last 6 months | Recent reports from this county mostly sit around that median. |
That's why the explorer shows them as separate layers you switch between, with different marks (a circle for crowd, a diamond for press) so the difference doesn't depend on color.
4. My speed test is far below the county median
The most common cause is obstruction: trees, roofs or poles blocking part of the dish's view of the sky. The Starlink app reports it as a percentage. Even a few percent can cut speeds and cause drop-outs, so tests with more than 1% obstruction are kept in the raw data but not counted in county medians. A congested county and a blocked dish look the same on a speed test; we only want to measure the first.
Other common causes: testing over weak Wi-Fi far from the router, other devices downloading during the test, and testing at peak evening hours (which is real and worth reporting, with the time band set correctly).
Fixing obstructions and installation is outside what this site covers. The Starlink app's obstruction check is the place to start.
5. Reading p25, median and p75
Fig.Hypothetical example, not real data
How to read a speed cell: p25, median, p75
Line up every accepted test from lowest to highest. The median is the middle one. A quarter of tests fall below p25, and a quarter are above p75. The shaded box is the middle half: "a typical test here lands somewhere in this range."
Illustration with made-up numbers · n = 15 · real cells publish only at n ≥ 20
Help your county get real numbers
Add it to the record. It takes about two minutes. We ask for your county, never your address, name or email.