Skip Navigation


Broadband speed test results for September 2015
Friday 02 October 2015 11:06:36 by Andrew Ferguson

The speed test results for September 2015 are in and speeds appear to have dipped for a good number of providers, though the overall UK picture shows a slight rise after a couple of months of dips. It is not easy to tell if this the actual user experience overall or whether we are seeing a shift in the demographics of those testing or possible a shift in what packages people are actually buying.

One theory for the decline in FTTC service speeds is that as the monthly cost decreases more people are upgrading even though they will not see the best speeds from FTTC or maybe are just opting for cheaper up to 38 Mbps rather than the premium up to 76 Mbps. To look at this aspect we will blog in more detail about speed profiles later in October.

One aspect that is changing is we are seeing more mobile, tablet and other devices testing and this means the negative effects of Wi-Fi on speeds will be having a greater effect. Given the growth in actual Internet use on these devices we believe it is fair to include these devices in the overall results, at the end of this figure laden article we do show some data on the differences between devices.

Large Provider Fibre Based Connection Speed Tests September 2015
Provider Median Download Mean Download Median Upload Mean Upload
FTTC Overall (excludes Virgin Media) 25.5 Mbps 27 Mbps 7.7 Mbps 8.4 Mbps
BT 28.4 Mbps 29.6 Mbps 8.5 Mbps 8.9 Mbps
Plusnet 30.8 Mbps 31.4 Mbps 9.1 Mbps 9.9 Mbps
Sky 21.4 Mbps 21.9 Mbps 7.3 Mbps 6.7 Mbps
TalkTalk 25.1 Mbps 25.8 Mbps 2 Mbps 4.6 Mbps
Virgin Media 39.8 Mbps 46.9 Mbps 5.1 Mbps 6.2 Mbps
ADSL/ADSL2+ Connection Speed Tests September 2015
Provider/Area Median Download Mean Download Median Upload Mean Upload
All Providers 4.9 Mbps 6.2 Mbps 0.8 Mbps 0.7 Mbps
BT 4.1 Mbps 5.6 Mbps 0.7 Mbps 0.7 Mbps
Plusnet 5 Mbps 6.2 Mbps 0.7 Mbps 0.7 Mbps
Sky 4.7 Mbps 6.1 Mbps 0.8 Mbps 0.7 Mbps
TalkTalk 5.2 Mbps 6.3 Mbps 0.7 Mbps 0.7 Mbps
Rural ADSL 3.1 Mbps 3.9 Mbps 0.4 Mbps 0.5 Mbps

The ADSL/ADSL2 services show the biggest change in the upload side, for BT and PlusNet this might be the result of the continuing roll-out of WBC ADSL2+ which is starting to become more available in areas such as the Highlands.

September was the first full month for tracking the new Vodafone Connect service and it sits pretty high up the table, and is a result of not having a legacy user base of ADSL/ADSL2+ customers, the profile of testers suggests three quarters are signing up to the FTTC services with perhaps 1 in 10 on the fastest up to 76 Mbps product.

The 20 Fastest UK Broadband Providers in September 2015
(ordered by median speed)
Smaller providers without enough geographic data samples are not included
Provider Median Download Mean Download Median Upload Mean Upload Download Speed of top 10%
Hyperoptic 66.6 Mbps 127 Mbps 48.8 Mbps 103.6 Mbps 398 Mbps
Virgin Media 41.4 Mbps 47.9 Mbps 5.1 Mbps 6.2 Mbps 100.7 Mbps
FastNet 22.2 Mbps 26.6 Mbps 7.7 Mbps 9.4 Mbps 51.3 Mbps
Vodafone Connect 20.7 Mbps 24.1 Mbps 8.7 Mbps 8.4 Mbps 42.6 Mbps
IDNet 18.9 Mbps 27.8 Mbps 4.5 Mbps 6.6 Mbps 67.4 Mbps
Relish 18.1 Mbps 18.5 Mbps 2.6 Mbps 3.3 Mbps 38.7 Mbps
AAISP 16.3 Mbps 28.4 Mbps 2.2 Mbps 6.2 Mbps 69.9 Mbps
Claranet SOHO 15.3 Mbps 25 Mbps 1.9 Mbps 11.9 Mbps 64.7 Mbps
EE Mobile 14.5 Mbps 18.2 Mbps 2.1 Mbps 5.3 Mbps 40 Mbps
XILO Communications 14.1 Mbps 25.4 Mbps 1.2 Mbps 7.4 Mbps 71.4 Mbps
BT 13.5 Mbps 19.6 Mbps 2.6 Mbps 5.4 Mbps 42.2 Mbps
Zen Internet 13.1 Mbps 22.1 Mbps 1.4 Mbps 6.7 Mbps 59.8 Mbps
Vodafone Mobile 10.8 Mbps 15.2 Mbps 2.5 Mbps 4.6 Mbps 36.2 Mbps
Entanet 10.5 Mbps 24.1 Mbps 1.6 Mbps 10.6 Mbps 63.1 Mbps
Plusnet 10.4 Mbps 17.8 Mbps 1.1 Mbps 4.9 Mbps 43.1 Mbps
Daisy 9.6 Mbps 17.3 Mbps 2.8 Mbps 6.3 Mbps 43.3 Mbps
Sky 8.4 Mbps 12.5 Mbps 1.1 Mbps 3.2 Mbps 30.5 Mbps
KC 7.7 Mbps 21.4 Mbps 0.8 Mbps 5.6 Mbps 61.6 Mbps
Tooway Direct 7.6 Mbps 8.7 Mbps 0.4 Mbps 0.6 Mbps 15.4 Mbps
TalkTalk 7.1 Mbps 12.6 Mbps 0.9 Mbps 1.9 Mbps 34 Mbps

Providers not making the top twenty this month with their median download speeds are EE 6.2 Mbps, Three 5.9 Mbps, 02 Mobile 5.5 Mbps, Eclipse 5.1 Mbps, Demon 4.8 Mbps, BT WiFi 3.8 Mbps and Post Office 2.6 Mbps.

Comparison of different hardware and its effect on speed test results
Median download speed
Provider Desktop/Laptop Tablet Mobile Phone
All Providers 14.2 Mbps 10.5 Mbps 11.6 Mbps
BT 14.4 Mbps 11.1 Mbps 14.2 Mbps
Plusnet 11.7 Mbps 7.8 Mbps 9.0 Mbps
Sky 9.3 Mbps 8.2 Mbps 8.0 Mbps
TalkTalk 7.7 Mbps 6.9 Mbps 7.5 Mbps
Virgin Media 44.6 Mbps 27.7 Mbps 33.7 Mbps

Looking at the overall picture, 65.5% of tests were carried out on a desktop/laptop, 16.8% on tablets and 17.7% on mobile phones. What is interesting is when we audit testing every day, we do see phones and tablets getting maximum speeds for FTTC and cable services so the fact the overall result is lower points to the problems and variability of Wi-Fi. Given the large amount of handheld device web browsing that goes on in 2015 excluding those devices from speed tests would remove the experience that one third of the population undergo.

If one was to try and pin down how fast broadband was in the UK on average the mean figure for desktop/laptop connections gives a mean download speed of 24.8 Mbps (5.7 Mbps upload) for September 2015. For those doubting the figures the speed profiles are available download and upload, the profiles illustrate nicely the various maximum speed points for the various products.

Comments

Posted by binary about 1 year ago
I wonder what TBB's thoughts are about how some providers seem to prioritise traffic to speed test sites compared to regular traffic.
Posted by andrew (Favicon staff member) about 1 year ago
If you want to email the names of the providers you believe are engaging in this practice can look for evidence in our testing.
Posted by gerarda about 1 year ago
Andrew How do tell for the purposes of these results whether a connection is by FTTC or ADSL?
Posted by keith969 about 1 year ago
This all goes to show that broadband is not a conventional service like e.g electricity where I can expect to get the same voltage wherever I live.

It would be great if I could get Hyperoptic or even Virgin Media but it's never going to happen here. So these comparisons are meaningless IMHO.
Posted by andrew (Favicon staff member) about 1 year ago
@gerarda Magic but having an availability checker of our own does help a lot http://labs.thinkbroadband.com/local/

And in A/B tests we've done we are more accurate than asking users themselves.
Posted by gerarda about 1 year ago
I was wondering because you don't ask on your test and so I did not see you knew whether my 1.5Mbps connection was FTTC or ADSL
Posted by TheEulerID about 1 year ago
Perhaps there ought to be different stats for results on FTTC that are below 40mbps and those above 40mbps. OK, not a perfect way of distinguishing between the two, but perhaps good enough to show a trend.

Posted by andrew (Favicon staff member) about 1 year ago
If the characteristics are the same for a location be it FTTC or ADSL, then probably in the ADSL figures.
Posted by andrew (Favicon staff member) about 1 year ago
@TheEulerID difficult to tell difference between the two products, especially with things like peoples home wiring and other variables.

Also some people order the 76/19 product for other reasons, e.g. package benefits, higher traffic priority but can only get under 40 Mbps speeds.
Posted by WWWombat about 1 year ago
@keith969
Actually, your electricity supply is indeed limited - not in voltage, but in current. Probably to 100 Amps.

If you tried to take up aluminium smelting as a hobby, you'd probably find yourself needing more than the cable could supply.

Gas and water are limited by the pressure, which limits the amount that can be supplied.
Posted by Blackmamba about 1 year ago
Hi Broadband Watchers.
When I ordered my FTTC BT checked my usage and advised me to take the lower option and has turned out OK 36 down and 8 up the only problem I had was using my I/Pad no 1 on Thinkbroadband speed which could not cope. My wife's I/pad 3 was OK but took the 180 days to get the correct results on the post code.
Posted by andrew (Favicon staff member) about 1 year ago
What do you mean took 180 days?

Speed tests to the map are updated over night, but you cannot stop others with services like ADSL/ADSL2+ from testing in the same postcode.
Posted by WWWombat about 1 year ago
Why are average FTTC speeds dropping?

The article posits this might mean takeup from more people with poor estimates ... but this ought to cause a reduction in upstream speeds too, which isn't happening to the same degree.

Likewise if more are choosing the cheaper packages. The 10 or 2 limit should be seen in upstream averages reducing too.

I wonder whether it is more a sign of crosstalk building. Perhaps the effects are more limited to downstream.
Posted by andrew (Favicon staff member) about 1 year ago
Overall we have seen FTTC upload speeds decrease too.

TalkTalk upload is stable, PlusNet dropping but ignoring as the change in upload speed on products will be impacting. BT dropping for a few months after a peak earlier in the year but levelled out almost now. Sky still dropping away.

Crosstalk is a possibility, as are lots of other things, so looking like a combination effect.

Based on forum experience, self-installs may be impacting too, as a fair few people get improvements when the faceplate, ring wire, extension issues are explained.
Posted by Blackmamba about 1 year ago
Hi Andrews Staff.
Could it be where the Post Codes were split and the results were on a differant location. I do not recall when this took place it may have been in the 180 day window I feel this could have produced peak and troughs. I have just checked to days generated results and the UK 24 % section has improved greatly over the last 2 months.
Posted by andrew (Favicon staff member) about 1 year ago
What UK 24% section.

I will also repeat that the coverage data is independent that we track is independent of speed tests that people run.

Posted by professor973 about 1 year ago
Something bent somewhere. Plusnet faster than BT with all the congestion problems they have been having. no surprise there then.
Posted by keith969 about 1 year ago
@ WWWombat

100amps is indeed the limit on the circuit breaker in my consumer unit, but I get that, guaranteed, there is no 'up to' figure. And I pay for what I use. Same with gas and water. With broadband however I have to pay a rate that does not give me the 'up to' rate quoted. So although I am paying for an 80/20 service, I only get 50/8. Yes I know the technical reasons but the cost model sucks.
Posted by andrew (Favicon staff member) about 1 year ago
So a better idea would be guaranteed speeds and charge per Gigabyte as with other utilities?
Posted by TheEulerID about 1 year ago
@andrew

Just about what I was to post. The equivalent of kWh is the GB. People pay for their electricity by the kWh, not by the current that flows.

It just happens that historical issues limit the achievable data rate. The UK's phone network was never designed with BB in mind.
Posted by keith969 about 1 year ago
@TheEulerID
Yep I fully agree. Charge by the Gb would be a much fairer way - people that stream videos should pay more than those who just use basic email / web page viewing.
Posted by chilting about 1 year ago
Paying by the Gigabyte certainly sounds attractive but it would be the absolute death blow to rural broadband.
With even less income being earned from slower rural lines BT would simply forget about rural users.
Posted by comnut about 1 year ago
The problem with rural broadband is the price of maintaining the lines... It is the 'BT monopoly' that does not help..
In America (hey please correct me!! :) ), the big monopolies were shut, so now they all have to *compete* for customers, giving great bargains and service - and their 'rural areas' are twice as big...
Posted by andrew (Favicon staff member) about 1 year ago
Masses of federal money going into rural america broadband initiatives, and as for choice lots more people have the choice of just one provider i.e. wholesale products are rare and you get towns where its cable or nothing.

FiOS where available is good, and other fibre initiatives are growing, but difficult to actually find info on premises passed/connected in many cases.
Posted by chilting about 1 year ago
@comnut
The problem with rural broadband is the price of maintaining the lines... It is the 'BT monopoly' that does not help..
The answer is simple - don't have fixed line broadband - don't use BT - use fixed wireless for the final 5%.
Posted by tommy45 about 1 year ago
@TheEulerID You say that it should be per GB used,would this also have line rental and subscription charges included ? also would BT start to offer SLA's for consumer BB where by any congestion or outages, and we could claim compensation for Because if you started to go down that route this is what people would expect
Posted by comnut about 1 year ago
A lot of peeps use 'pay as you go' phones, due to 1 month contract, no other charges.. Why cannot BT do that???
Posted by comnut about 1 year ago
I mean get a fixed line BB for 'pay as you go' NO other charges...
You must be logged in to post comments. Click here to login.