Home
 » ISP News » 
Sponsored

UPD Average UK Home Broadband ISP Speeds Climb to 17.8Mbps – Ofcom

Tuesday, April 15th, 2014 (10:11 am) - Score 2,139
fast-broadband-uk

The telecoms regulator has today published its latest biannual report into the real-world Internet speeds of fixed line broadband ISP consumers across the United Kingdom, which reveals that the average download rate has risen to 17.8Mbps (up by +17.41% from 14.7Mbps in August 2013). Sadly a huge gap still exists between urban and rural areas.

Overall Ofcom’s report, which is based on a fairly small sample of real-world data sources collected using custom SamKnows monitoring routers during November 2013 (installed in 2,391 homes around the country and resulting in 735 million separate tests), clearly shows that Internet connectivity performance is continuing to improve. Even the average upload speed managed to climb from 1.8Mbps six months ago to 2.3Mbps now.

Most of the recent improvements have come from the growing uptake of BT’s 80Mbps capable Fibre-to-the-Cabinet (FTTC) lines and Virgin Media’s DOCSIS cable platform, the latter of which recently announced another round of speed boosts up to 152Mbps (here). BT are also examining Vectoring technology for a possible deployment this year (here), which could improve their FTTC performance.

ofcom_average_uk_broadband_isp_speeds_by_technology_q1_2014

As usual the older / pure copper ADSL (up to 8Mbps) and ADSL2+ (up to 20/24Mbps) based broadband ISP lines haven’t shown much of an improvement, which is to be expected because in many cases they were already running at their best and doing so in a saturated market.

The deficiencies of ADSL have also contributed to the Digital Speed Divide between urban and rural areas, which exists because rural homes often connect via much longer copper lines and this slows the connection considerably (signal degradation). Rural areas are usually also the last to benefit from the latest NGA superfast broadband connections.

Ofcom’s “indicative analysis” shows that the average download speed in urban areas increased by 21% to 31.9Mbps in November 2013, almost three times the estimated average download speed in rural areas (11.3Mbps). By comparison the increase in download speeds for rural areas in the six months to November 2013 saw a change from 9.9Mbps to 11.3Mbps.

ofcom_average_uk_rural_vs_urban_broadband_speeds_q1_2014

The good news is that the Government’s Broadband Delivery UK (BDUK) programme is working to make fixed line superfast broadband services available to 95% of the country by 2017, which when combined with Virgin Media’s upgrades should ensure that performance and coverage of faster connectivity continues to rise for at least the next few years. But truly remote rural areas in the final 5% might well have to wait even longer for a solution.

Meanwhile the proportion of broadband connections classed as “superfast” (i.e. offering an “advertised speed” of 30Mbps+) is also on the increase and has now hit 25% of residential broadband connections (up from a figure of 19% in August 2013).

A Look at Specific ISP Performance

So what about the ISPs themselves, how do they all perform? Unfortunately Ofcom’s report is based on such a small sample size that it can only reflect the performance of the largest broadband providers, which is an important consideration because smaller ISPs tend to pay more attention to quality.

ofcom_average_download_speed_by_isp_q1_2014

Overall it’s no surprise to find that ADSL2+ providers typically exhibit fairly similar performance across the board because they all suffer from the same basic limitations, although KC does seem to deliver a slightly more consistent performance than the others.

Meanwhile, on the superfast broadband (30Mbps+) front, Virgin Media’s 120Mbps package (this has since become 152Mbps) can still churn out the fastest download speeds, which is because the top headline speed of BT’s closest FTTC technology is by comparison just 80Mbps (note: BT’s 330Mbps FTTP service doesn’t show due to limited uptake and availability). Never the less BT’s FTTC products can still hold their own, albeit only up to a point.

It’s also worth taking a quick look at the latency performance of ADSL2+ and superfast broadband services. Latency is a measure of the time (delay in milliseconds, 1000ms = 1 second) that it takes for a single packet of data to travel from your computer to a remote server/computer and then back again (ping). Lower is always better (faster) and this measure is especially important for fans of online multiplayer games, where a low ping gives you smoother and more accurate gameplay.

ofcom_average_latency_by_isp_q1_2014
ofcom_average_latency_by_superfast_isp_q1_2014

Generally speaking most of the ISPs, be they based on slower ADSL2+ or superfast (30Mbps+) technology, tend to exhibit similar performance levels. Meanwhile everybody is fairly level on the superfast front with little to tell them apart, although FTTC lines seem to show more variance.

Speed Complaints

So performance is improving, except if you’re stuck on one of those slower ADSL / ADSL2+ lines like most people, although occasionally people can experience significant problems with their speeds and when that happens it’s often important to be mindful of the relevant rules.

Firstly the Advertising Standards Authority, specifically their guidelines for tackling misleading promotions of “up to” broadband speeds (here), require ISPs to “demonstrate that [their] advertised speeds are achievable by at least 10% of users“. This is why so many providers now promote a “typical” speed (average performance based on that 10%) instead of the technologies top speed.

Meanwhile people who do suffer serious speed problems can sometimes get help through the regulators Voluntary Broadband Speeds Code of Practice (Version 2), which requires member ISPs (Listed Here) to explain to new customers the access line speed that they’re likely to achieve at home, and to try to resolve any problems when speeds fall significantly below the estimate. If the problem cannot be resolved then customers should be able to leave their ISP, without penalty, within the first 3 months of a new contract.

Ofcom are currently revising their speed code with some improvements for release later this year (here).

Ofcoms UK Fixed Line Broadband Speed Study (November 2013 Data)
http://stakeholders.ofcom.org.uk/../Fixed_bb_speeds_Nov_2013.pdf

UPDATE 12:42pm

The Government’s Communications Minister, Ed Vaizey, said: “Ofcom’s report confirms the remarkable transformation of UK Broadband currently underway. The UK has the best superfast coverage of all five leading European economies, and the news that average speeds continue to rise is tremendous news for homes and businesses alike. We are working hard to close the digital divide between urban and rural locations and are investing £790m to ensure that 95% of the UK will have access to superfast speeds by 2017.”

Leave a Comment
21 Responses
  1. Avatar gerarda says:

    Another piece of spin. A small self selecting sample, slow lines excluded, no definition of average but probably using the misleading mean average instead of median. I would imagine the mode average speed is probably zero but I suspect we will never see Ofcom publishing that.

    1. Avatar New_Londoner says:

      @Gerarda
      Unlike many of the reports of this type, Ofcom does in fact include details of its methodology, starting on page 58. You might find it informative to read this to find details of the sample size, statistical analysis etc.

    2. Avatar gerarda says:

      I had indeed only read the summary. However the statistical analysis proves they are using mean average and spends 12 pages trying to give a gloss of credibility to a measure they should never have used in the first place.

      My earlier statement stands.

  2. Avatar PhilB says:

    Its like say 10 people get 100mb and 90 get 2mb which means the average speed is 11.8mb.
    While they are pushing out 120mb+ there are hell of a lot of users still getting sub 2mb.No one is interested in those with 2mb,just keep pushing out mega fast speeds for a small minority.

    1. Avatar Steve Jones says:

      Indeed, Ofcom (and others) need to read some books on statistics. The arithmetic mean is wholly inappropriate for characteristing something, like downloads, with heavily skewed distributions. That’s why statisticians will prefer median, quartile, percentiles and the like in these circumstances (which is what tends to be used on income distribution, which has a similarly skewed distribution).

  3. I am one of the 2391 SamKnows monitoring sites and have been for several years. In January I switched from ADSL2+ @ 3.5/0.8Mb on long rural lines to FTTC @ 60.0/18.5Mb so my experience along with others making the same migration will be reflected in the improved Ofcom averages.

    1. Avatar gerarda says:

      If all the money being pumped in to FTTC was not improving the average speed something would be very wrong – however Ofcom’s method inflates the actual average speed and probably the absolute improvement over time. If we add your result to PhilB’s sample the average speed by Ofcom’s method would have increased when you moved to FTTC by 0.5mb even though 100 out the sample of 101 showed no improvement.

  4. Avatar Rowley says:

    In Kettering Northants on the aged copper wires and I get a whopping 1.5 MB. Just a tad short of the average 🙂

  5. Avatar No clue says:

    Wow look at those nasty maximum latency peaks on FTTC.

  6. Avatar MikeW says:

    To those arguing about “mean vs median”, I have a question:

    What do you think should be used in 2 years time, when take-up of superfast is expected to reach 50%? (superfast including cable and FTTx connections)

    At that point, the median will stop reflecting the better end of the ADSL market and start reflecting the slower end of the FTTx/cable market.

    Most people who complain about use of the mean, and suggest the median, do so because they feel it (the mean) fails to reflect the poor speeds on ADSL, and how vastly different they are from the NGA speeds.

    So, what (in 2 years time) would be a better way to show the statistics, when the median no longer reflects this either?

    1. Avatar gerarda says:

      The median always reflects mid point so is always better when describing the experience of the typical member of the population. This is why the ONS base average earnings on the mean.

    2. Avatar gerarda says:

      correction median not mean

    3. Avatar MikeW says:

      @gerarda
      I agree that the median does indeed usually manage to describe the typical experience of the typical member.

      However, that works best when the distribution of members is typical too – perhaps following a normal distribution of speeds.

      But when the distribution is skewed, with one big bunch of low speeds, and one big bunch of high speeds, then this “typical member” in the middle is no longer typical: He is either an outlier on the first bunch (atypically high speeds on ADSL) or he is an outlier on the second bunch (atypically low speeds on FTTx/cable).

      Looking at figure 1.2 of the PDF report shows the dilemma. The difference in height between the 3rd and 4th groupings is the problem, and the fact that there isn’t really a gradual transition from one to the other.

      Right now, the median reflects a reasonable “typical” measure of someone with decent ADSL2+ connection. But as take-up of NGA increases, that median gradually becomes someone on a stupendous ADSL2+ connection, and quickly changes to someone on a mediocre FTTx line or low cable package.

      I can’t help but feel that, in a year or so, the median will stop being particularly useful… and will continue that way for another couple of years, until it reflects a typical person within the FTTx/cable bunch.

      I’m not convinced that either the mean or the median actually reflect the take-up of broadband in the country. I think we need a different picture, but what beats the simplicity of a single number?

      It makes me think of Hans Rosling’s programme: The Joy of Stats. Perhaps we need an animation like his one on World Health.

    4. Avatar Steve Jones says:

      Personally I don’t like using a single number to characterise even moderately complex systems. I’ve had far too much experience of senior managers (and politicians) thinking it’s even possible. In the case of broadband, then simple speed distribution could be represented by deciles, quartiles or some such.

      Of course, all this misses the point that not everybody is interested (or needs) speed, and cost may be more important. Consequently, when somebody makes a rational economic decision, then they might well not adopt the fastest available speed. Where this choice is available, actual speeds may not reflect what is available.

      There seems to be this assumption that the faster average speed is always a proxy for overall economic benefit, where the reality is much more complex.

    5. Avatar gerarda says:

      The median is unlikely to jump suddenly as there will still be a number of sub 24mb FTTC users that will smooth things out. It will certainly not suffer from the distortion which even in the same report using the same data came out with figures of 17.8mb and 25.3mp as the average national speed

  7. Avatar Nick Lamb says:

    Actually the statistical ramp down for FTTC and FTTN is very steep as we’ve seen in previous reports. Using a median would cause apparently enormous rises in the “average” bandwidth as FTTC and FTTN went from 40% penetration to 60%. You can look at the bandwidth “spectrum” of a relatively small set of users and see a cluster instantly if they have (say) Virgin’s hybrid fibre/ cable FTTN, whereas with ADSL2+ and older technologies you get this long smear which is very noisy at the top end.

    There are two reasons. One is that the newer technologies are capped, so whereas essentially nobody got 24Mbps from a “up to 24Mbps” ADSL2+ service, lots of people are buying 40Mbps FTTC service and seeing 40Mbps. The other is that logically many more people are in the “sweet zone” where the technology performs best than could be true for exchange based systems. I’ve lived within spitting distance of a BT Exchange when ADSL originally rolled out, but there were probably less than 200 subscribers like me in the whole city (of six telephone exchanges) which means very few benefited fully. In contrast tens of thousands live within spitting distance of an FTTC enabled cabinet, because there are several hundred of those and they’re located (for obvious reasons) right in the middle of residential areas.

    So within the city medians will probably go from 8Mbps to 30+ Mbps almost overnight as uptake increases, some of those will be people at the edge of exchange districts going from a pitiful 1.3Mbps to 80Mbps because the FTTC cabinet is literally by their front window, and others will have gone from a perfectly respectable 18Mbps to 40Mbps because hey, a few pence per day extra for faster Internet, why not? But the headline “median” wouldn’t reflect these different experiences any more than a mean would. You need a spectrum chart of count versus bandwidth to show it properly. And you can’t put such a chart in a headline.

    1. Avatar gerarda says:

      But the median would still be the speed at which same number of people got a faster rate and those who got a slower rate, which is still a better measure than average which is only attainable by a minority

    2. Avatar Robert says:

      “.. lots of people are buying 40Mbps FTTC service and seeing 40Mbps.”

      An ISP only has to demonstrate 10% of its user base gets 40Mb to advertise it as an Upto 40Mb service as is mentioned in the news item.

      If “lots” of people were still getting 40Mb BT and many others would still be selling it as Upto 40Mb and not Upto 38Mb. So clearly lots of people DO NOT get 40Mb.

      Why is it regular people here with their essays can not read the news item?

  8. Avatar MikeW says:

    @Nick, @steve, @gerarda

    Nick requested a “spectrum chart” of count versus bandwidth, which I’d call a frequency distribution. I decided that this was a good idea – so created some. They are based on speedtest results on Thinkbroadband, gathered from some of their graphs in 2013 and 2014.

    UK Download Speed Distributions for FTTC and ADSL

    I present a breakdown of the speeds in the following ways:
    – ADSL speed distribution. The expected graph, skewed with many low speeds
    – FTTx 40Mbps distribution. Unexpectedly, it skews with many high speeds
    – FTTx 80Mbps distribution. More balanced, but still skews towards the higher speeds

    I’ve also provided a graph that combines all subscribers. With some old Sky data shown alongside some more recent BT data (who have the highest market share of FTTC subscribers), this graph gives you an idea of how the distribution is changing over time.

    I think these graphs highlight what I was saying: A “median” that sits somewhere in the 15-25Mbps region is atypical of the experience of either an ADSL subscriber or an FTTC subscriber.

    The two distributions for ADSL and FTTC-40 are really polar opposites, with skew in opposite directions.

    1. Avatar gerarda says:

      mikew

      Interesting analysis but see my reply to Nick. I think that is still better than a measure which ascribes 160 times more weighting to a 80mbps connection than a 0.5mbps one

  9. Avatar superslow says:

    Ofcom the monkey team how the hell do they know now that’s just a estimate as to me fttc is pathetic round our area most we can get is upto 20meg so I class that not superfast but superslow adsl speeds not fibre .

Comments are closed.

Comments RSS Feed

Javascript must be enabled to post (most browsers do this automatically)

Privacy Notice: Please note that news comments are anonymous, which means that we do NOT require you to enter any real personal details to post a message. By clicking to submit a post you agree to storing your comment content, display name, IP, email and / or website details in our database, for as long as the post remains live.

Only the submitted name and comment will be displayed in public, while the rest will be kept private (we will never share this outside of ISPreview, regardless of whether the data is real or fake). This comment system uses submitted IP, email and website address data to spot abuse and spammers. All data is transferred via an encrypted (https secure) session.

NOTE 1: Sometimes your comment might not appear immediately due to site cache (this is cleared every few hours) or it may be caught by automated moderation / anti-spam.

NOTE 2: Comments that break our rules, spam, troll or post via known fake IP/proxy servers may be blocked or removed.
Cheapest Superfast ISPs
  • Hyperoptic £22.00
    Avg. Speed 50Mbps, Unlimited
    Gift: None
  • Onestream £22.49 (*29.99)
    Avg. Speed 45Mbps, Unlimited
    Gift: None
  • xln telecom £22.74 (*47.94)
    Avg. Speed 66Mbps, Unlimited
    Gift: None
  • Plusnet £22.99 (*35.98)
    Avg. Speed 36Mbps, Unlimited
    Gift: £50 Reward Card
  • Vodafone £23.00
    Avg. Speed 35Mbps, Unlimited
    Gift: None
Prices inc. Line Rental | View All
The Top 20 Category Tags
  1. BT (2768)
  2. FTTP (2746)
  3. FTTC (1783)
  4. Building Digital UK (1740)
  5. Politics (1662)
  6. Openreach (1619)
  7. Business (1429)
  8. FTTH (1340)
  9. Statistics (1240)
  10. Mobile Broadband (1221)
  11. Fibre Optic (1062)
  12. 4G (1052)
  13. Wireless Internet (1020)
  14. Ofcom Regulation (1014)
  15. Virgin Media (1004)
  16. EE (696)
  17. Sky Broadband (668)
  18. Vodafone (666)
  19. TalkTalk (661)
  20. 5G (514)
Promotion
Helpful ISP Guides and Tips
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
»
Sponsored

Copyright © 1999 to Present - ISPreview.co.uk - All Rights Reserved - Terms , Privacy and Cookie Policy , Links , Website Rules , Contact