a a a Display Options Cymraeg
Follow Ofcom on Facebook Follow Ofcom on Twitter Subscribe to the Ofcom RSS Follow Ofcom on YouTube Follow Ofcom on YouTube

Independent regulator and competition authority
for the UK communications industries.

Search Ofcom


UK fixed-line broadband performance November 2011

Published 02|02|12

Overview of UK broadband speeds

UK broadband speeds increasing in 2011

1.1 Our research found that average actual UK fixed-line residential broadband speeds continued to increase during the second half of 2011, and in November 2011 the average actual speed was 7.6Mbit/s, a 0.8Mbit/s (11%) increase compared to the 6.8Mbit/s average recorded six months previously in May 2011 and 22% faster than the 6.2Mbit/s average recorded in November/December 2010 (Figure 1.1).

Figure 1.1 Average UK broadband speeds, November 2011

Source: SamKnows measurement data for all panel members with a connection in November 2011
Panel Base: 1,226

Notes: (1) Data have been weighted by ISP package and LLU/non-LLU connections, rural/urban, geographic market classification and distance from exchange to ensure that they are representative of UK residential broadband consumers as a whole; (2) As sufficient sample sizes were not available for consumers on packages of 'up to' 2Mbit/s or less, data collected for these packages in April 2009 has been factored in, in proportion to share of all connections in November 2011; (3) Data collected from multi-thread download speed tests. 1.2 While the average actual speeds recorded for 'up to' 8/10Mbit/s connections increased in the six months to November 2011 (up from 5.1Mbit/s to 5.3Mbit/s), the average speeds for 'up to' 20/24Mbit/s connections and above 'up to' 24Mbit/s connections fell, from 7.4Mbit/s to 7.0Mbit/s and from 38.2Mbit/s to 35.5Mbit/s respectively.

1.3 These changes in speed (the increase as well as the falls) can be largely attributed to the migration of BT and Virgin Media customers onto higher-speed services.

1.4 BT is currently upgrading its ADSL network from ADSL1 (which offers speeds of 'up to' 8Mbit/s) to ADSL2+ (which offers speeds of 'up to' 20Mbit/s) and is moving its customer base (and those connections which it provides to other ISPs on a wholesale basis) onto the faster services. As a result, a higher proportion of 'up to' 8/10Mbit/s connections are cable connections, which on average provide higher average download speeds than those provided using ADSL (see Figure 2.1), and the average speed for these connections has therefore increased.

1.5 There are two main factors which have contributed to the decrease in the average actual speed of 'up to' 20/24Mbit/s connections.

  • Firstly, BT has continued the upgrade of its network to offer ADSL2+. As BT's roll-out of higher-speed ADSL services covers more of the UK, it encompasses more rural areas, where the average length of the copper wire from the local exchange to the customer's home is longer. Because the speeds offered by DSL technologies decrease as the length of this line (the 'local loop') increases, the average speed offered by 'up to' 20/24Mbit/s ADSL services will fall, resulting in a decline in the average speed of all 'up to' 20/24Mbit/s services.
  • Secondly, Virgin Media launched an 'up to' 30Mbit/s cable broadband service in February 2011; in doing so, it withdrew its 'up to' 20Mbit/s cable service for new customers and allowed its existing 'up to' 20Mbit/s customer base to upgrade to the higher-speed service for a one-off payment of £30. As customers take advantage of this offer, the proportion of 'up to' 20/24Mbit/s connections provided using cable has fallen, and (for the reason outlined previously) so has the average speed of these connections.

1.6 The increasing take-up of Virgin Media's 'up to' 30Mbit/s service is the main driver behind the decline in average speeds for connections with a headline speed above 'up to' 24Mbit/s. As the proportion of above 'up to' 24Mbit/s connections that are 'up to' 30Mbit/s increases, there will be a fall in the average speeds provided by these connections, as they provide lower average speeds than the previously available connections in this category, which had either a headline speed of 'up to' 40Mbit/s for FTTC or 50Mbit/s for cable. (Virgin Media is also rolling out an 'up to' 100Mbit/s cable service, but this was not included in the research.)

1.7 Therefore the falling average speeds recorded for both 'up to' 20/24Mbit/s connections and above 'up to' 24Mbit/s connections are not a result of individual customers receiving lower speeds in fact the opposite is likely to be true as consumers are migrating to higher speed packages. Rather, it comes as a result of changes in the mix of connections within these two connection categories as new broadband products are launched. More than half of UK residential broadband connections have a headline speed above 10Mbit/s

1.8 The migration of UK residential broadband customers onto higher-speed packages means that more than half of all UK residential connections now have a headline speed of above 'up to' 10Mbit/s. Data provided to Ofcom by ISPs making up around 95% of the residential broadband market show that in November 2011, 58% of connections had a headline speed greater than 10Mbit/s, compared to 48% in May 2011 (Figure 1.2).

1.9 This migration to higher-speed packages is the main driver behind the increase in overall average UK broadband speeds and it is notable that these upgrades are often made at little or no additional cost to the customer. While most LLU-based services have already been upgraded to 'up to' 20/24Mbit/s ADSL2+-based services, BT is in the process of upgrading its ADSL network, and Virgin Media has increased the headline speed of its mid-tier offering from 20Mbit/s to 30Mbit/s; it also announced in January 2012 that it intends to double the headline speed for the majority of its connections by mid-2013.

1.10 The difference between the monthly rental fees for 'superfast' services (those with an 'up to' headline speed of 30Mbit/s or more) and 'current generation' services (which have a lower headline speed) is often relatively small. In January 2012 BT's 'up to 20Mbit/s ADSL More Broadband and Evening & Weekend Calls service had the same monthly line fee (£18 a month plus fixed phone line rental) as its More Broadband and Evening & Weekend Calls with superfast BT Infinity service which offered exactly the same bundled calls, download allowance and WiFi hot-spot use, but a faster 'up to' 40Mbit/s FTTC-based broadband connection.

1.11 As another example, based on tariffs available in January 2012, Virgin Media's 'up to' 30Mbit/s service typically cost £5 a month more than its basic 'up to' 10Mbit/s cable service, while TalkTalk charged £10 a month for the Fibre Optic Boost required to have an 'up to' 40Mbit/s FTTC connection rather than its ADSL2+ service (which is marketed as the fastest speed that the customer's line can handle). Plusnet's line rental and evening and weekend calls service cost £5 a month more with 'up to' 40Mbit/s rather than an ADSL2+ connection, although the latter had a higher download cap of 60GB (rather than 40GB) per month, and no annual contract (the price difference was over £13 if the ADSL service was taken with a lower 10GB a month usage allowance).

Figure 1.2 UK residential broadband connections by headline speed

Source: Ofcom, based on data provided by the UK's largest ISPs by retail market share (representing over 90% of the total market) Fibre-to-the-cabinet connections are getting faster

1.12 Analysis of the average download speed performance of UK broadband connections, by headline speed and technology, shows little change between May 2011 and November 2011 (Figure 1.3). The main difference was for 'up to' 40Mbit/s FTTC connections, where average speeds were 90% of the advertised speed in November 2011, up from 85% in May 2011 and 80% in November/December 2010.

1.13 In November 2011, ADSL-based connections continued to deliver average download speeds that were much lower than the headline 'up to' speeds which are frequently used to advertise broadband services. 'Up to' 8Mbit/s and 'up to' 20/24Mbit/s ADSL connections delivered just 41% and 31% of headline speeds during the period, in line with results from previous research, while cable and FTTC-based services on average delivered between 90% and 103% of headline speeds.

Figure 1.3 Average download speeds (24 hours) as a proportion of headline speeds by connection headline speed and technology

Source: SamKnows measurement data for all panel members with a connection in November 2011
Panel Base: 1,226
Notes: (1) Data have been weighted by ISP package and LLU/non-LLU connections, rural/urban, geographic market classification and distance from exchange to ensure that they are representative of UK residential broadband consumers as a whole; (2) Data collected from multi-thread download speed tests.

1.14 Analysis of the speeds recorded for 'up to' 40Mbit/s FTTC connections shows that while the 24 hour average increased by 5.0Mbit/s to 36.1Mbit/s between November/December 2010 and November 2011, the increase in maximum speeds (up 4.2Mbit/s) was less than that recorded at peak times (up 9.9Mbit/s) over the same period (Figure 1.4). This suggests that increasing FTTC speeds come as a result of BT having increased backhaul capacity in its core network to deal with the extra demand associated with superfast connections.

Figure 1.4 Maximum, average and peak download speeds for 'up to' 40Mbit/s FTTC connections, November/December 2010 to November 2011

Source: SamKnows measurement data for all panel members with a connection in each period 1.15 The increase in average FTTC connection speeds is also apparent in Figure 1.5, which shows that between November/December 2010 and November 2011 average download speeds increased by 4.2Mbit/s (13%) while at the same time advertised speeds were unchanged at 'up to' 40Mbit/s. BT announced in May 2011 that it would "roughly double" the speed of its FTTC service in 2012.

1.16 The average speed of connections provided using ADSL and cable technologies also increased over the period, although the main driver behind these increases was the shift of consumers onto higher-speed packages rather than improvements in the performance of existing services.

Figure 1.5 Average download speeds for fixed broadband connections, all connections including 'up to' 2Mbit/s and less, by technology

Source: SamKnows measurement data for all panel members with a connection in November 2011
Panel Base: 1,226
Notes: (1) Data have been weighted by ISP package and LLU/non-LLU connections, rural/urban, geographic market classification and distance from exchange to ensure that they are representative of UK residential broadband consumers as a whole; (2) Data collected from multi-thread download speed tests.

1.17 The significant differences between the advertised and actual download speeds of ADSL-based services led the Advertising Standards Authority to ask CAP and BCAP (the Advertising Code writing bodies) to conduct a review of the use of 'up to' speed claims in broadband advertising, in order to ensure that the advertisements do not mislead consumers.

1.18 Following this review, CAP and BCAP published guidance on the use of speed claims in broadband advertising in September 2011. This will come into force in April 2012 and requires that speed claims should be achievable by at least 10% of an ISP's customers and, where a significant proportion of customers do not receive a speed sufficiently close to the advertised speed, further qualifying information (for example, a statement that a proportion of customers achieve speeds within a certain range) should be included in the advertisement. Ofcom's response to the CAP/BCAP consultation had recommended that a 'Typical Speed Range' (TSR) should be used when advertising broadband on the basis of speeds, and that the TSR should have at least equal prominence to the maximum 'up to' speed (which should be achievable by a material number of customers).

1.19 Figure 1.6 below shows the average speed that at least 10% of the connections in our research received in November 2011, split by technology and headline speed, as well as the average speed that at least half of connections obtained, and an example of the 25% to 75% 'typical speed range'. For all of these different metrics, the figures for connections provided using cable and FTTC were much closer to the headline speeds than they were for ADSL connections.

Figure 1.6 Average speeds obtained by at least 10% and 50% of panellists and estimated typical speed ranges (25th to 75th percentile), by technology and connection headline speed

Source: Ofcom based on SamKnows measurement data for all panel members with a connection in November 2011
Note: Figures are rounded down to the nearest Mbit/s

1.20 In 2008 Ofcom and a number of ISPs agreed a Voluntary Code of Practice on Broadband Speeds. ISPs signing up to the Code committed to providing consumers with certain information at the point of sale, including an estimate of the maximum line speed (access line speed) they are likely to obtain. ISPs also committed to explain to customers that their actual broadband speeds are likely to vary for a host of different reasons, and provide information and advice on how consumers can improve their broadband performance.

1.21 With the co-operation and agreement of ISPs, an updated and strengthened Code came into force on 27 July 2011. This requires ISPs to give more accurate estimates of a customer's expected speed in the form of a likely range and ensures that customers' speed-related problems will be resolved by their ISP if possible, and if this is not possible, then customers whose speed is significantly less than expected have the ability to leave their provider within three months of the start of their contract without penalty. Most of the UK's largest ISPs are signatories to the Code.

Download speeds fell during peak times, but for some services more than others

1.22 Capacity constraints on ISPs' networks (contention) means that average download speeds frequently fall during peak periods when, typically, a larger number of connections are sharing the same bandwidth (Figure 1.7). While the average download speed recorded in November 2011 was 7.6Mbit/s, during the peak weekday hours of 8pm to 10pm (on average, the hours in the week when speeds were slowest) the average recorded speed was 7.4Mbit/s, 98% of the 24-hour average and 90% of the 8.2Mbit/s average maximum speed delivered (typically recorded during the 'off-peak' hours of 12am to 6am). Both of these proportions were unchanged since May 2011.

1.23 The relative performance of UK broadband connections varied by technology during the weekday 8pm to 10pm peak period in November 2011: 'up to' 8Mbit/s ADSL connections delivered just 84% of the 24-hour average at peak times, compared to 96% for 'up to' 30Mbit/s cable connections.

Figure 1.7 Variations in download speed by time of day, November 2011

Source: SamKnows measurement data for all panel members with a connection in November 2011.
Panel Base: 1,226
Notes: (1) Data have been weighted by ISP package and LLU/non-LLU connections, rural/urban, geographic market classification and distance from exchange to ensure that they are representative of UK residential broadband consumers as a whole; (2) As sufficient sample sizes were not available for consumers on packages of 'up to' 2Mbit/s or less, data collected for these packages in April 2009 has been factored in, in proportion to share of all connections in November 2011; (3) Data collected from multi-thread download speed tests.

In this section

Back to top