97% of Customers Not Getting Full Broadband Speed
The UK’s most comprehensive broadband speeds research reveals that the UK’s average actual fixed-line residential broadband speed has increased by over 25% over the past year from 4.1Mbit/s to 5.2Mbit/s as internet service providers (ISPs) increasingly move to offer higher speed broadband packages.
Ofcom’s broadband speeds research, conducted in partnership with broadband monitoring specialists SamKnows, shows that nearly a quarter (24%) of UK fixed-line residential broadband connections had a headline or advertised speed of above ‘up to’ 10Mbit/s in May 2010, compared to just 8% in April 2009.
However, the move to faster headline speeds has led to a growing gap between the actual speeds delivered and the speeds that some ISPs use to advertise their services. Differences between headline and actual speeds are often caused by broadband being delivered over copper lines which were originally designed for phone calls; hence speeds slow down over long and poor quality lines, and because of electrical interference.
|Headline (Advertised) speed||Average actual speed|
|‘Up to’ 8/10Mbit/s DSL packages*||3.3Mbit/s|
|‘Up to’ 20/24Mbit/s DSL packages||6.5Mbit/s|
|‘Up to’ 10Mbit/s cable services||8.7Mbit/s|
|‘Up to’ 20Mbit/s cable services||15.7Mbit/s|
In April 2009, average actual (or download) speeds were 4.1Mbit/s, 58% of average advertised ‘up to’ speeds (7.1Mbit/s). In May 2010, average download speeds were 5.2Mbit/s, 45% of average advertised ‘up to’ speeds (11.5Mbit/s).
The research found that cable broadband services delivered significantly faster actual (or download) speeds than comparable DSL services:
- Virgin Media’s ‘up to’ 10Mbit/s and ‘up to’ 20Mbit/s cable services delivered average download speeds around twice as fast as DSL packages with the same or similar headline speed (see table below).
- Virgin Media’s ‘up to’ 50Mbit/s cable service was the fastest service tested, delivering average download speeds of around 36Mbit/s with single thread tests and around 46Mbit/s with multi-thread tests conducted between 4 and 6am. A single thread test involves downloading a single file, while a multi-thread test involves conducting three file downloads simultaneously, which is more typical of the way heavy internet users consume broadband.
The impact of the length and quality of a consumer’s telephone line on DSL services led to wide variability in speeds received:
- Of panellists on ‘up to’ 20/24Mbit/s DSL services, almost two thirds (65%) had an average download speed of 8Mbit/s or lower, while only 2% received average download speeds of more than 14Mbit/s. No DSL panellists on ‘up to’ 20/24Mbit/s received average download speeds of more than 18Mbit/s.
- Of panellists on ‘up to’ 20Mbit/s cable services, 80% received average download speeds of more than 14Mbit/s and 4% received average download speeds of 8Mbit/s or below, with 19% receiving average download speeds of more than 18Mbit/s.
Actual speeds also slowed down in busy periods as heavier internet traffic leads to congestion on ISPs’ networks:
- The cable broadband service on average delivered higher download speeds at all times of the day than comparable DSL services, however it showed a greater slowdown during peak periods than some DSL providers.
|Average download speed during May 2010 (Mbit/s)|
|ISP package||24 hours
|4-6am every day
|‘Up to’ 8/10Mbit/s packages|
|AOL Broadband ‘up to’ 8Mbit/s+||3.6 to 4.7||3.4 to 4.4||4.0 to 5.2|
|BT ‘up to’ 8Mbit/s||3.8 to 4.5||3.4 to 4.1||4.0 to 4.7|
|O2/Be ‘up to’ 8Mbit/s||4.3 to 5.0||4.2 to 5.0||4.6 to 5.4|
|Orange ‘up to’8Mbit/s+||3.3 to 4.2||2.6 to 3.4||4.1 to 5.3|
|Plusnet ‘up to’ 8Mbit/s||3.3 to 4.2||3.0 to 3.9||3.4 to 4.4|
|Sky ‘up to’ 10Mbit/s||3.9 to 4.9||3.8 to 4.9||4.1 to 5.2|
|TalkTalk ‘up to’ 8Mbit/s||3.6 to 4.3||3.4 to 4.1||3.8 to 4.7|
|Virgin Media ‘up to’ 10Mbit/s||8.6 to 9.0||7.5 to 8.0||9.1 to 9.5|
|‘Up to’ 20/24 to ‘up to’ 50Mbit/s packages|
|BT ‘up to’ 20Mbit/s||6.1 to 7.6||5.6 to 7.0||6.4 to 8.1|
|O2/Be ‘up to’ 20/24Mbit/s++||8.1 to 9.7||7.9 to 9.4||8.9 to 10.8|
|Sky ‘up to’ 20Mbit/s||7.0 to 8.6||6.9 to 8.5||7.5 to 9.3|
|TalkTalk ‘up to’ 24Mbit/s||6.5 to 8.4||5.9 to 7.6||7.3 to 9.7|
|Virgin Media ‘up to’ 20Mbit/s||15.2 to 16.5||13.4 to 14.9||17.4 to 19|
|Virgin Media ‘up to’ 50Mbit/s||33.4 to 36.7||31.8 to 35.2||45.9 to 47.4|
Ranges published above represent the margin of error in terms of the 95% confidence interval around the mean (or average value) – see note 7; they do not represent the range of speeds delivered, which are much greater than the ranges shown above.
All data refers to single-file download tests (average speeds are higher using multi-thread tests), except for 4-6am every day.
+ Caution small sample under 50.
++ Caution normalisation less effective due to low number of respondents with line lengths above 2km.
Strengthening the Voluntary Code of Practice on Broadband Speeds
Ofcom has strengthened the Voluntary Code of Practice on Broadband Speeds (“the Code”) which it introduced in 2008. ISPs who sign up to the revised Code commit to give consumers a more accurate and consistent estimate of the maximum speed likely to be achievable on their line. ISPs also commit to help consumers improve their speeds and give consumers the option to leave their contracts early without penalty (within three months of signing up to the service) if they receive a maximum line speed that is significantly below the estimate they are given at point of sale and the ISP is unable to resolve the problem.
Advertising broadband speeds
Some ISPs have in the past year moved away from advertising their services on the basis of ‘up to’ headline speeds but others continue to do so. Ofcom has discussed this issue with the Advertising Standards Authority (ASA) and the Committee on Advertising Practice (CAP) who are currently undertaking a review in this area.
Ofcom has made two suggestions to the ASA and CAP. We have recommended that:
- Speeds should only be advertised if at least some consumers are actually able to achieve the advertised speeds.
- Those who advertise according to “up to” speeds, should also include a “typical speed range” (or TSR) based on a standard currency to be developed, similar to those in other industries (for example, APR in financial services, and MPG in motoring).
Advice for consumers on broadband speeds and services
Many consumers could take some relatively simple steps to improve their broadband performance. Ofcom provides advice for consumers on the factors to take into account when choosing a broadband provider and how to improve broadband speeds at home.
Ofcom’s updated consumer guides can be found here: http://consumers.ofcom.org.uk/guides/.
Ed Richards, Ofcom Chief Executive said:
“Ofcom’s research shows that average speeds have increased which is good news, but there is scope for a further step change in the quality of the UK communications infrastructure.”
“Actual speeds are often much lower than many of the advertised speeds which makes it essential that consumers are given information which is as accurate as possible at the point of sale; this is what the new Code is designed to deliver.”
“We are delighted that all major ISPs have signed up which is a major step forward.”
The research provides independent, robust data on the actual speeds that UK residential consumers are getting on average from the broadband services included in the research. Over 18 million separate service performance tests were carried out in over 1,500 homes during May 2010. The research sample allowed Ofcom to robustly compare the performance of those ISP packages with more than 250,000 residential connections (as well as Virgin Media’s ‘up to’ 50Mbit/s cable service, where the characteristics of its user base made it possible to recruit sufficient panellists).
The research is continuing, and Ofcom is planning to publish a report every six months, with the next report scheduled for January 2011. We are keen to also include smaller providers in the future research, if feasible, and would note that while at present the research only includes the larger operators, there are many other smaller operators available, many of which may match or even better the performance of some of the packages in the report.
The research collects performance data from a volunteer panel who have connected broadband monitoring equipment to their routers. Ofcom is looking for more volunteers to expand the research. For further details, see: www.samknows.com
The full research report, which includes a glossary of terms, can be found here:http://stakeholders.ofcom.org.uk/market-data-research/telecoms-research/broadband-speeds/broadband-speeds-2010/