Network upgrades boost average broadband speeds
15 August 2012
- Average UK broadband speed has increased to 9.0Mbit/s
- Consumers are benefitting from ISPs' network upgrades
UK consumers are achieving average residential broadband speeds of 9.0Mbit/s, driven by the move to new 'superfast' services, Ofcom research reveals. 
Ofcom's latest research into fixed-line residential broadband speeds shows that, in May 2012, the average actual UK speed was 9.0Mbit/s, which is 2½ times faster than the average speed of 3.6Mbit/s recorded in November 2008 when Ofcom first began its speeds research. 
Ofcom's research now includes some new 'superfast' packages, including Virgin Media's 'up to' 60Mbit/s service and BT's Infinity 2 'up to' 76Mbit/s service, the launches of which have contributed to the rise in average speeds. The continuing trend of increasing speeds recorded in the research confirms that consumer migration to faster services is gathering momentum.
Table One: Average actual UK fixed-line residential broadband speeds since November 2008 
|Date||Average actual UK fixed-line residential broadband speed|
While some consumers actively choose to upgrade to superfast broadband packages to achieve higher speeds, many are benefitting from improved speeds as a result of internet service providers' (ISPs') network upgrades, at little or no additional cost to consumers.
In May 2012, over two-thirds of UK fixed-line residential broadband users (68%) were on packages with advertised speeds above 'up to' 10Mbit/s, an increase from 48% in May 2011.
The proportion of broadband connections which are superfast (i.e. they have an advertised speed of 'up to' 30Mbit/s or above) has increased in recent months with the launch of new superfast packages. By May 2012, 8% of residential broadband connections were superfast, compared with 5% six months previously and 2% in May 2011.
Residential superfast broadband connections are also getting faster, with average speeds increasing from 35.5Mbit/s in November 2011 to 35.8Mbit/s in May 2012. 
Copper, cable and fibre performance
Of the different types of broadband technologies, Ofcom's research found that cable broadband connections generated the greatest increases in average speed in the six months to May 2012 – up by 3.6Mbit/s (26%) to 17.9Mbit/s.
Over the same period, average speeds delivered by ADSL broadband – a technology that delivers broadband over copper wires – increased by 10%, from 5.3Mbit/s to 5.9Mbit/s.
Average actual speeds recorded for fibre to the street cabinet (FTTC) connections, however, fell by 12% (from 36.0MBit/s to 31.6Mbit/s) in the six months to May 2012. 
Network upgrades to meet the need for speed
The noticeable overall improvement in speeds is, in particular, the result of ISPs upgrading their broadband networks.
BT's upgrade of its copper ADSL network, for example, has seen many customers moved from ADSL1 technology to the faster ADSL2+ technology, while BT's upgrade of its FTTC service has seen an 'up to' 76Mbit/s service launched alongside its 'up to' 38Mbit/s service.
Similarly, in February 2012, Virgin Media started to double the speeds of most of its broadband connections, increasing the top speed of its fastest package to 'up to' 120Mbit/s.
Comparisons between ISPs' download speeds
Of the 12 ISP packages included in the report, Virgin Media's 'up to' 100Mbit/s service was the fastest, with the research revealing average actual speeds of 88.3Mbit/s over a 24 hour period.
Of the other superfast packages included in the research, the average download speed on BT Infinity's 'up to' 76Mbit/s service was 58.5Mbit/s3, compared with Virgin Media's 'up to' 60Mbit/s at 55.9Mbit/s. BT's 'up to' 38MBit/s package achieved speeds of 32.2Mbit/s3 whilst Virgin Media's 'up to' 30Mbit/s service had average speeds of 30.1Mbit/s.
During busy peak periods, a higher proportion of Virgin Media cable customers experienced speeds of less than 90% of their average maximum speed, compared to BT Infinity fibre customers.
Table Two:Average download speeds by ISP package 
These ranges reflect the average speeds that would be achieved 95 times out of 100 if the exercise was repeated with different panellists. If the range of two operators overlap, then these operators offer comparable performance. These ranges are not a description of the range of speeds actually measured.
*Caution: Small sample size (<50)
Fastest upload speeds
Whilst download speed is the most important single measure in determining broadband performance, our research also considers other factors including upload speeds. Upload speeds can be an important consideration, especially for those consumers looking to share large files or use real-time video communications.
The research found that BT Infinity's 'up to' 76Mbit/s service delivered the highest upload speeds of all the packages, averaging 15.6Mbit/s. Among the ADSL2+ packages included in the research, O2/Be's 'up to' 20/24Mbit/s service provided the fastest average upload speeds at 1.1Mbit/s.
In April this year, guidance on the use of speed claims in broadband advertising published by the Committee of Advertising Practice (CAP) and the Broadcast Committee of Advertising Practice (BCAP), came into force.
The guidance states that advertised speed claims now have to be achievable by at least 10 per cent of the relevant ISP's customer base.
Many ISPs have therefore changed the way they advertise their broadband services. For example, ADSL2+ services, which were previously often promoted using the technology's maximum theoretical speed of 'up to' 24Mbit/s (which was rarely achieved in practice), are now frequently being advertised as 'up to' 16Mbit/s. Some ISPs have moved away from promoting their services primarily on the basis of speed focussing instead on price, or added value features such as free security.
Ofcom broadband speeds research
This is Ofcom's seventh report into fixed-line residential broadband speeds using data collected by research partner SamKnows. It is the final report in this research period (May 2010 to May 2012). The next research phase will start in or around November 2012 for two years, and Ofcom will select a research partner in autumn 2012 following a tender process.
The report provides data on the average performance of fixed-line residential broadband in the UK and of the individual ISP packages included in the research. In total, these packages accounted for over half of UK residential broadband connections in May 2012.
The results help consumers to understand the significant variations in the performance of ISP packages and, when considered alongside other factors such as price, can help them to make more informed purchasing decisions.
In July last year, a revised and strengthened Voluntary Code of Practice on broadband speeds came into force. This requires ISPs to give more accurate estimates of a prospective customer's expected maximum speed in the form of a range. In addition, it seeks to ensure that, where possible, customers' speed-related problems will be resolved by their ISP.
If this is not possible, then customers whose speed is significantly below the estimated access line speed range have the ability to leave their provider within three months of the start of their contract without penalty. All of the UK's largest ISPs are now signatories to the Code.
Ed Richards, Ofcom Chief Executive, said: "Our research shows that the move to faster broadband services is gathering momentum. Consumers are benefitting from network upgrades and the launch of new superfast packages, giving them faster speeds and greater choice.
"We are continuing to work with the advertising code-writing bodies and ISPs to ensure that speeds advertised reflect actual speeds experienced, to allow consumers the ability to make informed decisions when shopping around to find the most suitable package."
NOTES FOR EDITORS
The research looked at 12 packages provided by the eight largest ISPs in the market by subscriber numbers. 559 million separate test results were recorded out in 1,831 homes in May 2012.
1. A number of BT Infinity panellists' measurement units reported anomalous results during the May 2012 reporting month, and for 22 out of a total of 239 BT 'up to' 38Mbit/s and 'up to' 76Mbit/s connections the downstream speeds recorded fell to less than 1Mbit/s at certain times. For more information on these issues please refer to page 5 of the report. The average UK residential actual broadband speed in May 2012 was 9.1Mbit/s when the data from those BT panellists whose measurement units reported anomalous results is excluded from the analysis.
2. Prior to November/December 2010, speed measurements were carried out using "single-thread" tests, which involve downloading one data file. From November/December 2010 onwards "multi-thread" tests were used to reflect the fact that ADSL2+ technologies were configured to handle multiple users/processes at the same time. However, as ADSL1 technologies provide similar results for both test methodologies and before 2010, most consumers were on ADSL lines, broad comparisons of national average speeds are possible between different time periods.
3. When data from those BT panellists whose measurement units reported anomalous results is excluded from the analysis, for May 2012:
a. The average actual speed of residential UK superfast broadband connections rises from 35.8Mbit/s to 36.9Mbit/s
b. The average actual speed of residential FTTC connections rises from 31.6Mbit/s to 34.0Mbit/s
c. the average download speed on BT Infinity's 'up to' 76Mbit/s rises from 58.5Mbit/s to 60.7Mbit/s
d. the average download speed on BT Infinity's 'up to' 38Mbit/s rises from 32.2Mbit/s to 34.7Mbit/s
4. Part of this fall was due to an issue affecting some panellists' BT Home Hub routers (see note 1) that constrained speeds over wired connections during the testing period. BT has been remotely updating these Home Hub routers to resolve this issue and has informed us that this has now been completed. Excluding the results from these panellists, the fall in average FTTC speeds was lower at 2.0Mbit/s (6%), with the average speed recorded in May 2012 being 34.0Mbit/s.
Also this comparison should be treated with some caution as the May 2012 data is the first time that the average FTTC speed figure has included connections provided on a retail basis by ISPs other than BT.
It is possible that the drop in speeds might also be explained by the continuing rollout of FTTC services to less densely populated areas. Although fibre to the street cabinet delivers much faster broadband services, as with ADSL technology, the greater the distance between the home and the street cabinet, the slower the speed.
5. Source: SamKnows measurement data for all panel members with a connection in May 2012
Notes: (1) Includes only ADSL customers within 5km of the exchange and in geographic Markets 2 and 3 and in the Kingston-upon-Hull area for Karoo; (2) Includes on-net customers only for LLU operators (3) Data for ADSL operators have been weighted to ISP regional coverage of LLU lines and distance from exchange; data for Virgin Media's cable service have been weighted to regional coverage only; (4) Data collected from multi-thread download speed tests; (5) The range shown represents a 95% confidence interval around the mean.(6) ADSL2+ technology offers maximum theoretical speeds of 24Mbit/s. Those providers who mention speed in advertising of these packages frequently promote speeds of 'up to' 16Mbit/s.
6. Other measures which Ofcom has taken to improve the information available to consumers choosing broadband services include publishing research on the quality of customer services, publishing data about the level of complaints received for ISPs and providing an accreditation scheme for price comparison sites. Ofcom has also published a consumer guide to improving broadband speeds.
7. The research was conducted in partnership with broadband monitoring company SamKnows. For more details, visit www.samknows.com.
8. The Communications Act 2003 requires Ofcom to make arrangements to find out about consumers' experiences of electronic communications services, which includes broadband services, and the way they are provided (section 14). We do this by carrying out research into their experiences of these services. Subject to certain exceptions, we have a duty to publish the results of our research and to take account of it in carrying out our functions (section 15).
9. Ofcom is the independent regulator and competition authority for the UK communications industries, with responsibilities across television, radio, telecommunications, wireless communications and postal services.