Broadband News

Peak and off peak performance at the largest UK broadband providers in January 2018

The comparisons between the different download tests our broadband speed test runs are important for a number of reasons and the simplest is that the download test of a single data stream is akin to an activity people engage in a lot and that is streaming films and TV shows to their various devices from the big screen TV to their mobile. Another issue is that the ISP wide statistics highlight the reality that lots of people are not buying the fastest services available.

So what do the figures for January 2018 tell us, well for most providers nothing dramatic. For Virgin Media the download speeds continue to increase, but we are also seeing a higher level of variability in individuals speeds as reflected by the quality metric.

Off-Peak Tests Results January 2018
Off-Peak defined as midnight to 5.59pm
Median Average
Providertbbx1 Test
(1 download)
httpx6 Test
(6 downloads)
% difference
between
tbbx1 and
httpx6 speeds
Upload SpeedQuality
Lower is Better
Grade A = Best
Latency
BT 28.9 Mbps 32.7 Mbps -11.7% 6.3 Mbps 0.4 - Grade A 39ms
EE 12.3 Mbps 14.9 Mbps -17.4% 1 Mbps 0.9 - Grade B 51ms
Plusnet 20.5 Mbps 24.4 Mbps -15.9% 4.3 Mbps 0.4 - Grade A 44ms
Sky 14.4 Mbps 16.5 Mbps -12.7% 3.1 Mbps 0.6 - Grade A 51ms
TalkTalk 13.7 Mbps 17 Mbps -19.4% 2.7 Mbps 0.5 - Grade A 52ms
Virgin Media 53 Mbps 73.9 Mbps -28.3% 6.1 Mbps 1.1 - Grade B 39ms

The changes in how broadband speeds are shown in broadband adverts and package listings is creeping every closer, and we will be shortly publishing a product based analysis of the average speeds between 8pm and 10pm for the larger providers. The providers are clearly going to use the multiple download test results from any system they use, and with Virgin Media increasing the over provisioning factor it is possible that many adverts and claims will not actually change that much.

Peak Tests Results January 2018
Peak time defined as 6pm to 11:59pm
Median Average
Providertbbx1 Test
(1 download)
httpx6 Test
(6 downloads)
% difference
between tbbx1 and
httpx6 speeds
Upload SpeedQuality
Lower is Better
Grade A = Best
Latency
BT 27.3 Mbps 31.2 Mbps -12.5% 6.1 Mbps 0.5 - Grade A 40ms
EE 12.1 Mbps 14.5 Mbps -16.6%

1.9 Mbps

0.8 - Grade B 54ms
Plusnet 21.3 Mbps 25.7 Mbps -17.1% 4.8 Mbps 0.4 - Grade A 42ms
Sky 13.7 Mbps 16.7 Mbps -18.0% 3.6 Mbps 0.6 - Grade A

50ms

TalkTalk 12.4 Mbps 15.3 Mbps -19.0% 2.4 Mbps 0.5 - Grade A 52ms
Virgin Media 48.6 Mbps 71.1 Mbps -31.7% 6.0 Mbps 1.3 - Grade C

41ms

For those wanting a broader explanation on the quality metric check our our speed test FAQ.

The difference between the httpx6 tests at off-peak and peak times are not that different, and but as we cannot guarantee that the same spread of connection types and geography are covered we don't make a direct comparison in this analysis of our speed test results. What you should compare is the % difference column and they are surprisingly similar in most cases, and there in lies a problem with the ASA system, i.e. while a median rather than a top 10 speed is more representative it is still not the speed that any individual will see unless they happen to sit on that median point. Providers have not unduly fought the change in speed rules and it is possible the marketing teams believe it may drive more people to buy the higher speed packages when people actually get around to switching providers.

Comments

Lol - interesting point on the potential for upselling to higher speed packages. The law of unintended consequences at work again?

  • WWWombat
  • 10 months ago

I wonder to what extent people do actually buy on advertised bandwidth rather their functional experience. There are plenty that post here that do, but I tend to think that's the exception. My deep suspicion is that most people will be content as long as their web experience remains reasonably snappy and their on-line video streams aren't buffering and so on. For many there was a qualitative difference when they went from ADSL to VDSL. It's less obvious they will want to trade up just because some headline number changes.

  • TheEulerID
  • 10 months ago

Post a comment

Login Register