Broadband News

Late or still on time? Superfast Cymru in the spotlight

The SuperfastCymru project along with many other projects has had a confused history, as all too often targets are mis-represented and while the dates for the coverage figures are open to debate as they have often been vaguely described the target for the Welsh broadband project is 96% with access to something fibre based, i.e. VDSL2 at any speed or FTTP as an option. This is different to the superfast coverage figures, which have never been released for Wales, but we are expecting that once the 96% fibre figure is hit, that superfast coverage will be 90%.

On the date side, there are some references to an original date of end of 2015, some newer references to end of 2016 and when the additional 42,000 premises where added in an extension a date of 2017 appeared. Our alarm clock is set to declare project failed if the 96% figure is not reached by 31st December 2016.

The Daily Post has referenced our figures for Wales and we compared these with the official output back in March and we have updated our summary to show the change since March. We generally expect our figures to undershoot compared to the official figures as we take a pessimistic view of the impact of cross-talk. It should be highlighted that there was an expectation that the SuperfastCymru project would deploy around 3-4% of premises with native FTTP, if the dates have been missed it may be the slower roll-out that FTTP usually has that is partly responsible for any slippage.

thinkbroadband calculation of Superfast, USC, USO and Fibre Broadband Coverage across the Wales
thinkbroadband (tbb) figures include gap funded and commercial
(figures in brackets are change since March 2016)
(*) areas where no commercial deployment existed before start of SuperfastCymru project
Area % fibre based % superfast
24 Mbps or faster
% superfast
30 Mbps or faster
% Openreach FTTP % Under 2 Mbps USC % Under proposed 10 Mbps USO
Combined Welsh Total 92% (+1.4) 87.7% 85% (+1.6) 0.56% (+0.16) 0.9% (=) 8.3% (-0.8)
Blaenau Gwent (*) 99.7% (+0.6) 96.6% 96.7% (+1.3) 0.82% 0.3% 0.9%
Bridgend 96.7% (+0.2) 93.7% 92.1% (+0.1) 0.1% 0.4% 2.3%
Caerphilly 97.5% (+0.3) 95.1% 93.3% (+0.5) 0.02% 0.4% 1.6%
Cardiff 97.6% (+0.3) 96.7% 96.3% (+0.9) 0% 0% 1.1%
Carmarthenshire 80.8% (+5.5) 73.2% 71.5% (+3.8) 0.01% 2.2% 18.3%
Ceredigion (*) 69.4% (+3.5) 55.9% 54.1% (+1.7) 0.82% 4.5% 34.3%
Conwy (*) 91.6% (+1.7) 86.1% 85% (+3.9) 0% 0.9% 8.9%
Denbighshire 83.1% (+0.4) 78.9% 77.9% (+1.5) 0.03% 0.6% 12.7%
Flintshire 93.1% (+0.2) 87.7% 85.9% (+0.3) 3.35% 0.5% 7.3%
Gwynedd (*) 85.5% (+0.7) 76.3% 74.1% (+0.5) 3.43% 2.1% 16.4%
Anglesey (*) 86% (+0.9) 78.3% 75% (+1.6) 4.97% 1.6% 15.7%
Merthyr Tydfil (*) 99.4% (=) 94.4% 93.6% (=) 2.25% 1.1% 2.5%
Monmouthshire 85.7% (+3.4) 76.2% 74.7% (+2.9) 1.65% 2.7% 14.2%
Neath Port Talbot 95.7% (+1.3) 92.4% 91.4% (+1.4) 0% 0.4% 3%
Pembrokeshire (*) 82.9% (+4.2) 74.8% 73.2% (+3.4) 0.4% 1.8% 17.3%
Powys (*) 70.2% (+5.7) 58.3% 56.9% (+4.4) 1.2% 3.5% 30.5%
Rhondda Cynon Taf 98.7% (+0.5) 96.3% 94.8% (+1.7) 0.07% 0.4% 1.5%
Swansea 95.5% (+0.8) 93.7% 93.3% (0.4) 0.39% 0.1% 2.2%
Torfaen 96.6% (+0.4) 95.2% 94.1% (+0.7) 0.35% 0.2% 1.8%
Vale of Glamorgan 94.7% (+0.9) 91.6% 89.4% (+1.5) 0.21% 1% 4.8%
Wrexham 92.5% (+0.4) 86.3% 84.4% (+0.1) 0.99% 0.7% 6.3%

So Wales is four percentage points short of the magic 96% figure, but the rate of change in two months and also over a longer period suggests that if the pace of delivery is maintained that 96% by 31st December 2016 is possible. Of course this is still no comfort to those totally missed out, and for those in VDSL2 areas who in the non-superfast and estimates suggest no improvement over ADSL/ADSL2+ the roll-out will be very frustrating.

On the figures the observant will notice that the superfast increase is larger than the fibre increase in some areas, this is because we continually are checking different parts of the UK to verify and where needed improve the model. People actually speed testing on the cabinets and in FTTP areas being a key factor in helping us to identify where the automated model needs some tweaking. We could hide these variances by only publishing the data once a year but believe much better to give the public as much information as possible as soon as possible.


The impressive bit is those where there was no commercial deployment before. Even the very rural areas of Powys and Ceredigion are close to 70% and Blaenau at 99.6% amy actually hit the 100% Fibre based. Its 0.9% under 10Mb in one of the most depressed parts of the UK may actually show whether economic activity will improve with better broadband provision! ( Or just be used for streaming video).

  • jumpmum
  • over 2 years ago

I am unsure how widespread this is, but in a fair number of part of rural Ceredigion the infrastructure is there, the web sites (including this one) say you can order and have done so for some month but BT, Zen etc won't take orders from a significant number of post codes. I don't know to what extent this causes the numbers to be overstated in the most rural areas, but probably makes little difference to overall average for Wales. It does make a difference to those effected.

  • Llety
  • over 2 years ago

@Llety example postcode(s). For some cabinets, especially ones that are for old EO areas there can be problems and we have pushed to get a few areas fixed. So given details of someone trying to order we can pester.

If we have cabinets as live and they are not, happy to remove them, and this has happened a few times where cabs have gone live and hit a snag for the first couple of orders and withdrawn.

  • andrew
  • thinkbroadband staff
  • over 2 years ago

I'm with @jumpmum - the change for Blaenau and Merthyr is staggering. To go from zero to 99% fibre is pretty impressive.

Places that have reached this kind of coverage have normally started with significant commercial coverage, including VM.

  • WWWombat
  • over 2 years ago

As usual, people ought to distinguish between what might be said earlier on by politicians as aspirational statements and what gets written down in contracts and put into actual project delivery plans (which often get tweaked as plans and aims get readjusted a little).

Very often when people talk about a project failing to deliver on time it's measured against those political statements. Fair enough to take politicians to task, but maybe not the project.

  • TheEulerID
  • over 2 years ago

The biggest source of delays, now as in the past, seems to be the EU.

  • WWWombat
  • over 2 years ago

I would estimate the Vale of Glamorgan to have a Openreach FTTP percentage of 2.34%+ by March 2017, going by the postcode areas that are either waiting for the FTTP infrastructure to be commissioned or having the FTTP infrastructure currently being built.

  • cymru123
  • over 2 years ago

In Pembrokeshire some Exchanges have constantly had the availability pushed-back in the case of my own, from May 2015 to September 2017. Who knows if that will be met?

Even if the Exchange has the capability, there's virtually no guarantee that the service some\all\any subscribers.

  • wittgenfrog
  • over 2 years ago

My neighbour is offered VDSL2 at up to 15Mbps. I am offered nothing. (I checked both line numbers on BT wholesale). Our wires go down same poles 1500m to cabinet 9 at Llanidloes. We have similar ADSL speeds. What is going on? How can I get this remedied? Asking BT retail has been useless so far. Can andrew enlighten?

  • gtrman
  • over 2 years ago

I am surprised to read that Bridgend shows as 93.7% with 24 Mb or faster and 96.7% fibre. The geography of the valleys is mostly single roads with perhaps a small number of side roads and cabinets often at one end ( as in my case ) leading to those at the other end ( as in my case ) getting sub 24Mb speeds.

Enter cf32 8pf in your broadband map, it shows only 5 speed tests and I think that only the 16.16 test is fibre, the cabinets (7) are located near enough at at the pin.

  • lmschuffer
  • over 2 years ago

Work your way north and you have to go a long way to find the next fibre speed test, in fact there might only be 6 in the whole valley.

Better brains than mine will no doubt make better sense of the figures here.

  • lmschuffer
  • over 2 years ago

On that exchange, it seems all 7 cabs are converted (so 100% fibre). 5 of them look to have almost all their properties (2,100) within 500-600m. Easily superfast.

Cab 6 and 7 (500 properties) are the exceptions, and have some longer lines. I'd guess at 150-200 on lines too long for superfast.

Result: About 7% too long. About right.

Speedtests can be poor judges of availability; if no-one buys or no-one knows TBB. And, of course, people can run tests on poor wifi, or poor internal wiring - so one-off tests don't help much unless you know the context.

  • WWWombat
  • over 2 years ago

WWWombat. I think your comment "if no-one buys or no-one knows TBB" makes my point. I thought the figures were "calculated" using TBB speed tests which don't appear to be supported by the online tests shown, unless older test are not on the map ?, where do the speeds come from ?, otherwise, is it a guesstimate albeit maybe a good one ?. Clearly there are more connections than the map tests shows.

  • lmschuffer
  • over 2 years ago

"I thought the figures were "calculated" using TBB speed tests"

Which figures were you thinking of?

The speeds shown on the map do indeed come from the TBB speed tests, so are actual speeds seen by someone - whether on a good or bad line, good or bad wifi, expensive package or cheap package.

Old tests are dropped from the map (after 6 mths?).

  • WWWombat
  • over 2 years ago

However, the coverage percentages shown in the table in this article aren't calculated from the TBB speed tests at all. They are calculated by figuring out the availability of services to different homes, taking into account line distances etc, and using their own (pessimistic) conversion of length to speed.

I don't know what TBB does to calculate the speed estimates, but it definitely doesn't use the numbers from the speed tests.

  • WWWombat
  • over 2 years ago

Post a comment

Login Register