Broadband News

UK Broadband Speed Test Results for January 2017

January brings about some changes to our speed test results, we have integrated our quality metric that assesses each and every test that we see and the aggregate result for a provider should help to reveal how good their service beyond the shouting about the best download and upload speeds. The quality metric is connection agnostic, so there is no reason a good 4G connection should not score as good as a pure fibre one if the quality measured over the duration of the test was as good. A score of 1 is the ideal, 1.2 to 1.4 covers the main bulk of providers and any provider above 3 is performing very badly.

The 30 Fastest UK Broadband Providers and Services in January 2017
(ordered by median speed)
Smaller providers without enough geographic data samples are not included
Provider Quality Metric Download Speed of bottom 10%
(Mbps)
Median Download
(Mbps)
Mean Download
(Mbps)
Median Upload
(Mbps)
Mean Upload
(Mbps)
Download Speed of top 10%
Mbps
B4RN 1.4 5.6 103.5 173.3 78.7 150.9 411.3
KCom Lightstream 1.3 22.5 55.3 77.4 12.4 19.1 160.4
AAISP (FTTC) 1 14.2 52.4 47.2 12.9 16.2 72.7
BT Infinity (FTTH) 1.3 21.2 50.2 63.7 11.2 14.2 129.4
Hyperoptic 1.5 18.3 49.9 76.2 35 56.4 113.7
Virgin Media 1.8 8.9 43.2 52 5.3 6.4 104.9
Gigaclear 1.3 15.3 42.1 64.3 40.1 67.6 120.5
Zen Internet (FTTC) 1.1 16 37.7 41.7 10.6 12.1 71.2
SeeTheLight - IFNL (FTTH) 1.4 16 36 40.1 11.6 17.4 63.2
Daisy Wholesale (FTTC) 1.4 10.2 34.4 35.7 9 11.2 67.8
BT Infinity (FTTC) 1.3 14.3 34.3 35.5 8.5 8.9 56.7
PlusNet (FTTC) 1.3 14.2 32.8 34.7 4.6 7.2 61.3
Entanet (FTTC) 1.4 8.8 31.3 32.8 11.8 12.3 61.8
Vodafone Broadband (FTTC) 1.4 12.8 30.7 32.3 8.7 9.6 57.5
Eclipse (FTTC) 1.1 9.5 29.6 32.8 6.9 9.5 69
TalkTalk (FTTC) 1.4 13.4 26.7 28 1.8 4.3 39.2
EE (FTTC) 2.5 11.3 26.7 26.6 8.4 7.8 37.4
Sky (FTTC) 1.4 11 24.9 25.4 7.1 7 37.3
Ask4(FTTH) 1.4 3.4 23.8 35.6 8.4 12.4 95.3
EE Mobile 2 3.2 20.8 28.6 3 5.5 61.4
Vodafone Mobile 2.3 3.3 14 16.8 2.3 3.7 32.7
O2 WiFi 2.3 3.1 11.8 14.3 1.6 3.4 29
O2 Mobile 2.5 2.5 11.5 15.7 2.1 3.5 34.4
Relish 1.9 2.4 10.1 15.8 1.1 1.9 35.1
Three Mobile 2.2 2.2 8.7 15.6 1.6 4.1 40.2
KCom (ADSL) 1.5 1.7 7 7.1 0.6 0.6 12.6
Zen Internet (ADSL) 1.4 1.3 6.3 7.1 0.6 0.6 15
BT WiFi 2.2 1 6 8.1 1.2 2.5 18.3
Demon (ADSL) 1.4 0.9 5.8 6.5 0.6 0.5 12.9
PlusNet (ADSL) 1.4 1.1 5.7 6.9 0.6 0.6 14.7
IDNet is missing as the sample sizes were too small to make the table

The quality metric should also help to inform in terms of its trend over time and to help with that in future charts we will indicate if it has gone up or down compared to the previous set of results.

Since its not possible to list all the providers, we have our table showing the results from the various technology groups. Of course things like the range of product speeds, quality of wireless from the home gateway a provider suppliers and other factors do come into play.

Speeds of Broadband Technologies as seen in January 2017
(ordered by median speed)
Provider Quality Metric Ratio of tests Down Speed of bottom 10%
(Mbps)
Median
(Mbps)
Mean
(Mbps)
Median Upload
(Mbps)
Mean Upload
(Mbps)
Down Speed of top 10%
(Mbps)
FTTH/FTTP 1.4 1.2% 18 51.4 77 17.6 33.6 168.1
Cable 1.8 22% 8.5 42.6 51.4 5.3 6.3 104.3
VDSL2/FTTC 1.4 36.5% 12.7 29.2 31.2 7.1 7.6 51.8
Mobile 2.2 N/A (*) 2.8 14.3 21.2 2.1 4.6 45.9
Fixed Wireless 2.0 0.16% 2.8 12.6 23.2 2.3 9.4 48.3
Satellite 1.8 0.03% 2.2 8.4 9.1 0.1 0.3 19.8
ADSL/ADSL2+ 1.5 40.2% 1.1 5.4 6.6 0.6 0.6 14.5

(*) Mobile was not included in the proportion of tests as the assumption is the majority of the tests are people who have a fixed line option too.

The peak versus off-peak battle continues and the Virgin Media DOCSIS service was worse in January 2017 than December 2016 suggesting the operator still has not got to grips with its various congestion hot spots and that number may be growing. It is interesting that the difference in the quality metric between day time and evening is significantly difference for Virgin Media too.

Median Peak and Off-Peak Download Speed Tests Results January 2017
Provider tbbx1 Test
(1 download)
httpx6 Test
(6 downloads)
Quality
  7am-3pm 6pm-midnight % difference 7am-3pm 6pm-midnight % difference 7am-3pm 6pm-midnight
BT 22.1 Mbps 24.3 Mbps +9.9% 25.8 Mbps 27.9 Mbps +8.1% 1.11 1.21
EE 8.4 Mbps 8.1 Mbps -3.6% 9.3 Mbps 8.9 Mbps -4.4% 1.44 1.73
Plusnet 19.2 Mbps 15.6 Mbps -18.8% 23.4 Mbps 17.2 Mbps -26.5% 1.12 1.17
Sky 10.7 Mbps 10.0 Mbps % 12.9 Mbps 11.8 Mbps % 1.28 1.35
TalkTalk 10.7 Mbps 9.9 Mbps % 12.1 Mbps 11.8 Mbps % 1.20 1.26
Virgin Media 40.3 Mbps 22.7 Mbps -43.7% 59.7 Mbps 41.8 Mbps -30% 1.59 1.91

G.fast is emerging now from its small scale trials and hopefully by April/May of this year we will be able to start looking at the performance and look at deriving a speed versus distance model based on observations from our data.

Comments

It's interesting that the likes of Gigaclear/Hyperoptic do not seem to perform as well as BT Infinity's FTTH. It would seem like actual performance does not reflect connection speed.

  • AndyCZ
  • 9 months ago

Home gateway and wireless speeds, plus product choice e.g. popularity of 20/1 product on Hyperoptic. Difference is apparent when you look at B4RN with single 1 Gigabit product.

  • andrew
  • thinkbroadband staff
  • 9 months ago

Andrew I think the Ratio of tests and the Quality metric columns are reversed in the 2nd table.

  • jumpmum
  • 9 months ago

So they were, have flipped the column headers around now.

  • andrew
  • thinkbroadband staff
  • 9 months ago

Have I missed something, because nowhere can I fina out what this "quality metric" is actually measuring in technical terms. Is it consistency of speed, or of latency, or packet loss rates, of single vs multi thread, or a combination of some and or things I haven't even thought of?

As it is, I really don't know what it is actually measuring.

  • TheEulerID
  • 9 months ago

Ouch at PlusNet. That may be a regional or gateway specific thing as I haven't see it myself.

  • AndrueC
  • 9 months ago

@ andrew - Wireless does make it tricky, but I think it's safe to say a relatively small number of Hyperoptic/Gigaclear customers take the top package. This does reenforce the argument that the demand for top speed FTTP packages is just not there at the present moment.

  • AndyCZ
  • 9 months ago

The joy of as soon as I explain things a week or two later someone else is doing the same

  • andrew
  • thinkbroadband staff
  • 9 months ago

I'm surprised that BT ADSL/ADSL2+ doesn't figure in your statistics. This technology has the 40% of the tests, so surely a sizeable proportion must be sold by BT?

  • Murgatroyd
  • 9 months ago

The main table is the 30 fastest providers/services and the lack of BT ADSL in that table tells you they are not in the top 30

Table was extended from 25 to 30 to allow some ADSL to be included, and general tech table will of course feature them

  • andrew
  • thinkbroadband staff
  • 9 months ago

I glad to see this I am moving for BT to Zen with hardly a difference in price! Zen does very well in Which? reviews which scores it highly for its customer services. Why the switch because one reason is their router is AC rather than BT's offering of only N. Bt for us has been absolute headache!! Its reliability has been suspect too.

  • ethicalme
  • 9 months ago

@AndyCZ the lack of demand for the top speed packages is a function of our market as much as anything else. Everyone say thank you to TalkTalk and Sky for treating broadband as a value add / freebie.

  • CarlThomas
  • 9 months ago

BT Infinity (FTTC) 1.3 14.3 34.3 35.5 8.5 8.9 56.7
The above from the first table seems odd to me because I always get about 75Mbps down on tests and 17Mbps up.
Surely there are a lot of users getting the same.?
Or am I not reading it correctly.?

  • ironman12345
  • 9 months ago

Only 3% of BT Infinity (FTTC) users get above 70 Mbps

11% get over 17 Mbps upload speeds.

  • andrew
  • thinkbroadband staff
  • 9 months ago

I'm in the 3%? (80/20 sync I get 74 down)

and 18 up so in the 11?

I thought it would be more than 3 and 11% overall.

  • ZenUser27
  • 8 months ago

Also can the stats above be trusted? after all your tester is not reliable.

http://www.thinkbroadband.com/speedtest/results.html?id=14863311375191421169

wth!? on other sites 74/18

Never seems to be right on your site - even when I do testing side by side

  • ZenUser27
  • 8 months ago

Only 3% of BT Infinity 80/20 customers get above 70 Mbps Andrew , that is not a lot unless I am doing something that the other 97% are not doing.?
Please do not take offence as none is intended , just seems odd to me.

  • ironman12345
  • 8 months ago

@ZenUser27 Yes it can be trusted, perhaps the others are rejecting some of the slow samples and then taking an average, In fact if I use the same algorithm as some others then we'd show

72 Mbps down and 18.4 Mbps for that test.

A higher number does not immediately mean it is correct.

  • andrew
  • thinkbroadband staff
  • 8 months ago

@irconman12345 Remember that is 3% of those on VDSL2 with BT Consumer, a rough split on product suggests this is 10% of the Infinity2 VDSL2 users and so in line with the advertising rule.

Be interesting to see any provider show 10% of customers getting more than 76 Mbps speed tests, given the 76 Mbps is maximum IP throughput.

  • andrew
  • thinkbroadband staff
  • 8 months ago

How about listing the slowest 30 providers, with their very worst speed tests. Some of those in your 'top 30' will, in my experience, figure in the bottom 30.

  • M100
  • 8 months ago

The slowest 10% figure is included to highlight the range of speeds, and if we thought people would be genuinely interested in a 50+ row table each month then adding more is just a longer to produce item problem.

  • andrew
  • thinkbroadband staff
  • 8 months ago

You may find that a lot of people only come here when they are having problems. If your connection is stable then most people don't bother testing.
This would sway the results in a big way.
It may be an idea to include a tick box for wether your connection has issues or not!

  • CodeBusters
  • 8 months ago

Could you write up an explanation of the "quality metric"? Something like "1 is perfect. Above 2 implies that online FPS gaming will have sporadic issues. Above 5, and you should expect IPTV to be erratic. Above 10 and the connection is completely unusable part of the time."

Basically something to tie the metric back to my experience as a customer.

  • farnz
  • 8 months ago

@Codebusters We run speed test services for a number of sites and thus the report includes all those too, not just those who seek out thinkbroadband when they have issues.

  • andrew
  • thinkbroadband staff
  • 8 months ago

On the quality metric - three ways of looking at connections without recourse to the actual download or upload speeds is talked about at http://blog.thinkbroadband.com/2017/02/speed-test-quality-metrics/

  • andrew
  • thinkbroadband staff
  • 8 months ago

Post a comment

Login Register