How we use estimated and actual speeds
As of November 2020, if a customer has an existing FTTC or GFAST service and uses their telephone number to check broadband availability, their actual speed will be used.
A series of calculations will then be made to determine the speed range & Stay Fast Guarantee. For more information, please visit Ofcom's website.
For customers with copper connections, those using a postcode to check broadband speed, or those currently without a broadband service, we will continue to use Openreach estimates. These estimates are based on the performance of similar broadband connections across the country.
How we calculate throughput overheads
We conform to the high-level testing principles set out in Annex 5 of the Voluntary Code of Practice.
- We partner with SamKnows to provide speed tests from our Home Hubs.
- We've built a test panel across our main products. This represents our product base across a range of exchanges, network status and line speeds.
- We run tests over 35K lines each month and update our results quarterly.
- We run tests throughout the day. Peak time is 8 pm to 10 pm.
- If we detect more than 160Kbps of traffic on the line, we won't run the test. This reduces the impact on our customers and avoids negatively impacting the results
- The speed test runs a 3 Thread TCP test to assess the speed of the broadband connection.
- We compare the actual throughput result to the sync speed at the time of the test. This helps us determine the normal and maximum overheads for xDSL products. We can then remove them from the relevant sync speed estimates.
- We use the 20th and 80th percentiles of the FTTP test panel's average speeds to calculate the point-of-sale speeds for FTTP products.
- Speed tests are conducted on servers located off the BT network, with no favourable traffic routing applied.