How we measure internet performance

For each destination, we use fixed-broadband tiles within a 10-kilometer radius of the city. The result summarizes the tiles’ average download speeds while giving more influence to tiles supported by more tests.

The data window is January 1 through March 31, 2026, covering Q1 2026. The observation date is March 31, 2026. The August 1, 2026 review date records the editorial review and does not change when the underlying tests occurred.

1. Select valid Speedtest fixed-broadband tiles within 10 kilometers of the destination.
2. Exclude tile observations without a valid average download speed or positive test count.
3. Order the remaining tiles by their average download speed in megabits per second.
4. Calculate the weighted median of tile average download speed, using each tile’s test count as its weight.
5. Display the result rounded to the nearest whole megabit per second.

The weighted median is the tile speed where cumulative test-count weight reaches at least half of the total eligible test-count weight.

The resulting metric score receives its descriptive rating through the common Retirement Lists rating bands. That rating applies only to Internet performance. It is not an overall Lifestyle score and does not imply that every address can obtain the measured speed.

Compare may emphasize an Internet difference only when the two displayed results differ by both:

• at least 25 Mbps;
• at least 20% relative to the lower displayed result.
• at least 20% relative to the lower displayed result.

This is a user-interface comparison rule. It does not represent measurement accuracy, statistical significance, or a confidence interval.

The measure fails closed when no valid fixed-broadband tile and test evidence is available within the required radius and data window. In that situation, Retirement Lists publishes no Internet result.

The system must not substitute:

• another city;
• a regional or national value;
• M-Lab observations;
• provider availability or advertised speeds;
• inferred performance; or
• manually estimated values.

Speedtest observations come from voluntary tests and describe measured fixed-broadband download performance during the stated quarter. They do not establish:

• broadband availability at every address;
• which providers serve a particular property;
• available plans or prices;
• long-term reliability;
• upload performance;
• latency;
• performance at different times of day; or
• a guaranteed address-level connection speed.

The 10-kilometer radius may cross municipal boundaries. Results should therefore be read as evidence about measured performance near the destination, not as a boundary-perfect municipal statistic.