Slowest street in the UK, or just bad statistics?
According to uSwitch in a PDF based press release they have identified the fifty slowest streets in the UK, based on over 2 million speedtests carried out over a few months, and apparently a street can only appear if 30 registered tests are carried out.
This press release is a periodical one, as we know the regional press eats up material like this, where they can talk about how bad broadband is in the area, but with some basic checks we are not at all sure about the actual data behind the uSwitch claims.
Cromarty Road, Stamford which is now the slowest in the UK at 0.132 Mbps has at least one postocde and an example one is PE9 2TQ for those that want to look on our map or uSwitch's own. Looking at the uSwitch map we can find the actual test which appears to have been carried out on 13th May 2012 @ 6:28pm using a BT connection, this appears to be the solitary speed test on the street, certainly no results from other providers are showing up, and makes us conclude that the 30 registered tests per street is not real (or they don't show all the data on their map) - looking at satellite imagery shows the road has around 24 properties in it. The BT Wholesale speed checker also indicates an ADSL2+ service of over 10 Mbps for the postcode, and other properties in the area are showing significantly higher speeds, and even FTTC is available in some streets nearby.
A little more checking reveals that when you do a comparison for Cromarty Road, uSwitch lists BT as the best provider at 15.2 Mbps, but interesting this is based on 434 tests in the exchange area, where the fastest was 121.91 Mbps. The total number of tests uSwitch list for the major providers is 1,370 for the exchange area, and given the size of Stamford we question the 30 tests in each street statement.
The UK like all other countries has varying speeds in different parts of it, and even if full fibre were installed to every property at the cost of billions, speed test results would still show variations down to things like wireless network congestion, malware on a computer, running a speedtest when already using the connection and the small matter than no residential connection is supplied as a guaranteed 1:1 contention service.
Obviously we run our own speedtests, and are not questioning that on 13th May someone in Cromarty Road did not get that speed, but we are doubtful of the analysis that labels this as the worst street for broadband in the UK. Alas in the rush for bad news headlines, and the time pressures faced by many press outlets and the lack of deep understanding of the technical issues, no-one will be doing any looking into whether the results make any sense.