Broadband News

Ofcom publishes Connected Nations 2016 report

The UK telecoms regulator Ofcom has released its analysis of UK broadband coverage and speeds for 2016, and there are some headline grabbing facts, such as 'Digital divide declines, but 1.4 million homes cannot get decent broadband'. The 1.4 million homes statement is referring to the number of premises who only have access to a fixed line broadband service of under 10 Mbps earlier in 2016.

Another headline that is likely to be seen a lot is that the UK has superfast broadband available to 89% of premises (30 Mbps and faster definition). The important part here is that this is a figure from May/June 2016 and rolling back our data it seems we are in agreement. Ofcom for 2016 is apparently working at the premises level, i.e. dealing with a dataset of 29 million premises, where we with a lot less resources available are working at the postcode level, so to be in such close agreement supports work we did back in 2011/2012 that showed that once you get above exchange sized groups of households that using the majority result for a postcode was reliable enough for tracking what the BDUK process was or was not delivering. There are some places in the Ofcom report where 12% are said to not have access to superfast broadband, which is at odds with their declaration of 89% superfast coverage, the presumption is that rounding errors have come into effect, which is a shame as if working to the premises level declaring percentage figures to several decimal places should be easy.

So back to the 1.4 million premises needing USO work, and it is now significantly smaller at 1.1 million premises and for decision makers this change and the pace of it is important, since delivering a solution to less premises becomes easier and worries about cost constraints might lead to the adoption of a poor solution to the problem. Ofcom does project some estimates, i.e. under 10 Mbps will drop to 4% by end of 2017 and 1% in the early 2020s.

Ofcom details three scenarios for the Universal Service Obligation.

  • Scenario 1, 1.1 million premises affected and based on not getting 10 Mbps from fixed line services. Any solution delivering 10 Mbps suitable with an estimated cost of £1.1 billion. Ofcom estimates this will be 0.3 million premises in early 2020s.
  • Scenario 2, 2.6 million premises affected, 10 Mbps download speed and a upload speed of 1 Mbps, medium latency (this appears to mean satellite is excluded), 50:1 contention ratio and 100 GB monthly usage allowance. Ofcom estimates the number in this scenario will be 2% in the early 2020s.
  • Scenario 3, 3.5 million premises, i.e. deliver a superfast solution of 30 Mbps and faster and uploads of 6 Mbps. Unlimited data allowance and committed information rate of 10 Mbps. Based on existing coverage a cost of around £2 billion is estimated. Ofcom estimates that in the early 2020s this will be 4% or 1.1 million premises

With Ofcom modelling 4% of the UK not having a superfast connection option in the early 2020's, this suggests the regulator believes that only 1% extra coverage will be delivered post 2017. The USO report does consider how peoples needs will change and that 10 Mbps may well not be enough in 2020, and dare we say it but if superfast roll-outs deliver the 95% in just over 12 months time and clawback work and commercial infill in urban areas continues and can push into the 97% to 98% region that making the USO a superfast one seems the way to go, only using lesser solutions for those outside cost thresholds.

For those having their works Christmas party tonight, have a glass of champagne as the UK average broadband speed has changed 'The average download speed of all active connections in the UK is now 37Mbit/s, an increase of 28% from 29Mbit/s in 2015'. Alas what is missing in many places where Ofcom use this figure is that this is the sync speed as supplied by broadband providers and thus with providers like Virgin Media over provisioning by 10% e.g. 100 Mbps service connects at 110 Mbps and the impact the transport layers have on IP throughput plus IP Profiling on BT Wholesale connections it is no surprise that we report lower speeds, our modelled speed for the UK is 25.7 Mbps and observed is 23.9 Mbps (modelled figures excludes mobiles and tablets to reduce impact of Wi-Fi and corrects for market share of different providers). Where we do agree closely is that if everyone bought the fastest connection available that the averages would be a lot higher, Ofcom states around 132 Mbps and we are declaring 129.8 Mbps. What we do on a monthly basis is report on the performance of ultrafast pure fibre (FTTH/FTTP) and availability of FTTH and FTTP increases expect to be able to talk about more providers.

"It is not clear from this data why some consumers are choosing not to take superfast services where they are available and further market research is needed. In order for superfast broadband take up to continue growing at the same rate as it has in the past, providers may need to consider new approaches for attracting customers, including articulating its benefits more clearly and lowering prices further.

Ofcom on superfast takeup

The Ofcom report acknowledges the increasing level of superfast take-up but suggests the rate of take-up may be slowing down and that prices may need to lower even further. Balancing this against the Governments, Ofcom and the peoples desire for a much higher availability of fibre to the premises across the UK is going to be a difficult task.

Comments

I can give good reasons why superfast takeup is low -- COST and LACK OF ISPs offering the product.
People do not want to pay for the speed and don't want to change ISP either.

  • gsmlnx
  • about 1 year ago

23 home ISPs and 71 business ISPs sell superfast.

  • Somerset
  • about 1 year ago

Not all ISPs are available when FTTC goes live. It is usual for those offering it to existing clients to have no real special deals to sign up and if you paid your line rental in advance, then you are stuck with existing provider as no refund will be given if you decide to switch. Often, only 24 month contracts are being offered as you have to take out a new contract to upgrade and that is a prime reason why not more people are switching. If you could simply upgrade an existing contract with no connection fee then the take up would be greater.

  • jabrady10
  • about 1 year ago

Is it me, or is just too convenient for Ofcom and the government that this report is published less than 24 hours after a debate about broadband in the House of Commons?

Too easy to allow Ministers to shrug off useful debate until 2017

  • 961a
  • about 1 year ago

Which providers force people with ADSL to take a 24 month contract when upgrading to VDSL2?

  • andrew
  • thinkbroadband staff
  • about 1 year ago

@Somerset
My comments where more on the FTTP/H lines so perhaps I should have used Ultrafast.
My neighbours can either have 2.2Mbps ADSL or FTTP but until very recently the vast majority have stayed on ADSL. And it has been available since 2011.
So very slow to change.
I would also be intrigued to see how many have chosen the bottom tier of 55/10 over the 330/30.

  • gsmlnx
  • about 1 year ago

@andrew - Slight URL error in the first link.

https://labs.thinkbroadband.com.local

  • Diggory
  • about 1 year ago

Sorry - second link, not first.

  • Diggory
  • about 1 year ago

Thanks on URL have fixed now pesky . slipping in when should have been /

  • andrew
  • thinkbroadband staff
  • about 1 year ago

Demand for ultrafast might not be there yet. For a home user 4K streaming sets the highest demand and that should work with 25M or so. For the vast majority not streaming 4K, a fast ADSL is enough and there is no reason to upgrade. And those with slow ADSLs generally can't upgrade as they are EO or very remote.

  • hvis42
  • about 1 year ago

The singular of premises is premises. The word premise means something completely different.

  • TheEulerID
  • about 1 year ago

Am I the only one who did not understand very much about these "scenarios"?

It is a jumble of premises, percents, Mbps and billions. What can someone with a 3 Mb/s line in London expect from them?

  • hvis42
  • about 1 year ago

Issue is the Openreach price is now so low that reducing it doesn't affect retail prices at all. Without regulation on retail providers there is really no way for OFCOM to reduce the price to consumers. If people don't want to pay any more for a faster service OFCOM will have to accept it. A lot of people can do everything they need with 15Mb. Just us technophiles that want ever faster but cheaper.

  • jumpmum
  • about 1 year ago

On the scenarios what you'd get would all depend on which scenario was adopted, and the cost to deliver something to you. They explore various thresholds, i.e. not what you pay but what provider would spend to deliver to you, e.g. £5,000 to deliver which at wholesale prices of £10/m would take a long time to recoup.

  • andrew
  • thinkbroadband staff
  • about 1 year ago

No commercial network operator would ever recover £5k at £10 a month. To do so would imply that there were zero operating costs and that the cost of capital was less than 2.4% PA. That's fantasy land.

The only way that it can be paid for is via a subsidy. That means a cross-subsidy (as rural phone lines are covered), by a state subsidy, by an industry levy or an operator being forced to take a loss.

  • TheEulerID
  • about 1 year ago

@TheEulerID - If your referring to FTTP, the lower packages are charged a one off activation fee, where as the higher packages are not charged that fee, two of my neighbours chose the 55/10 route due to if it was too slow they could always pick the next one up.

Also the FTTP packages are +18 Month minimal contract.

  • PaulKirby
  • about 1 year ago

@hvis42
4k streaming: Amazon Prime in 4k currently consumes around 16Mbps, into a Fire TV box. Not sure what codec is being used.

  • WWWombat
  • 12 months ago

@hvis42
On the scenarios...

Perhaps the point missed is that the Ofcom document is written to offer advice to the government, answering questions relating to "decent broadband" that the gov asked earlier in the year.

In answer, Ofcom use the 3 scenarios to define "decent":

Scenario 1 is to deliver a "standard broadband service" which has the only requirement being 10Mbps down.

Scenario 2 is "higher spec standard broadband" with extra requirements on upstream, latency, contention and monthly usage.

Scenario 3 is "superfast broadband", with even higher specifications.

  • WWWombat
  • 12 months ago

The addition of extra requirements in scenarios 2 and 3 mean that more premises need to be subsidised in each case, and the subsidy to upgrade each one increases.

  • WWWombat
  • 12 months ago

@gsmlnx
It is generally true that most subscribers stay on the lowest packages on FTTP, even when better is available. Certainly KCOM reported that for Hull.

I *think* we can see it in the Ofcom reports too, but I need to sift the data better.
...

  • WWWombat
  • 12 months ago

Australia's NBN, as public-run, provides great statistics on the takeup - including the breakdown within each technology type.

There, 25/5 has largest takeup at 45%, then 12/1 at 35%. Next, 100/40 gets less than 15%. Then 50/20 at 5%.

The 3 packages that are faster than 100/40 account for less than 100 subscribers, out of nearly a million. 0.01%

article: goo.gl/6xn5vv

  • WWWombat
  • 12 months ago

Sorry. Those "takeup" percentages are really a market share percentage.

  • WWWombat
  • 12 months ago

Sorry 2: And the market shares are specific to FTTP subscribers.

FTTN is still smaller there, but it follows the same basic pattern.

  • WWWombat
  • 12 months ago

You also can see the splits on products from things like the speed profiles we post on twitter @thinkbroadband The capped speeds of product tiers are usually very visible.

  • andrew
  • thinkbroadband staff
  • 12 months ago

Hi Fast.
The meeting was at Denbies wire estate post code RH56AA North of Dorking it was invertation only approx 100 people's attended SCC staff Openreach Bill Murray at the end a list of all the post codes were give out.
They did answer many questions over the roll out from the floor there were also set meetings at various Halls throughout Surrey advertised via Website which was organised by Katie.

  • Blackmamba
  • 12 months ago

@andrew
It's been a while since I've seen a roundup of the speed profiles. I'm thinking of things like the VM, BT, Sky, KC profiles at the bottom of this article:

http://blog.thinkbroadband.com/2014/11/october-2014-speed-tests-a-full-round-up/

Or, even better, the down & up profiles in this one:

http://blog.thinkbroadband.com/2014/12/detailed-analysis-of-november-2014-speed-tests/

  • WWWombat
  • 12 months ago

@PaulKirby

I was responding to Andrew's point that it would take a long time to recover a £5k investment if the wholesale price was £10 a month. That's to do with Ofcom's proposed threshold for capex before excess construction costs were to be charged. My point is it would never pay back at that rate.

As far as the £10 per month goes, that's about the cost of the 80/20 package. Lower speeds are more like £7.

  • TheEulerID
  • 12 months ago

The cost of providing decent speeds to many on long lines could be only a few hundred pounds each if openreach was prepared to use repeaters, as used in other countries.

  • brianhe
  • 12 months ago

Repeaters cause big problems with the ANFP unless every pair in a cable goes through a repeater at the same point. Then there is the little issue of power. Technically it's much better to use FTTrN if power is available.

  • TheEulerID
  • 12 months ago

Repeaters cause big problems with the ANFP unless every pair in a cable goes through a repeater at the same point. Then there is the little issue of power. Technically it's much better to use FTTrN if power is available.

  • TheEulerID
  • 12 months ago

@Euler
I think you are getting the £5k wrong.

In the case of a USO, it isn't meant to represent a commercially-viable amount which needs to justify payback. It is the point where (commercially-viable-amount + internal-USO-slush-fund-amount + external-USO-slush-fund-amount) runs out, and needs a top-up from the subscriber.

Today's telephony USO only expects money from an internal-USO-slush-fund of KC and BT.
...

  • WWWombat
  • 12 months ago

BT are currently offering a broadband USO which also implies an internal-USO-slush-fund too, provided LR-VDSL gets the OK.

Other proposals include an industry-funded-external-USO-slush-fund. Or perhaps a taxpayer-funded one.

Whatever other funds exist, you mustn't forget to add them.

  • WWWombat
  • 12 months ago

Learn your maths please.

There are some places in the Ofcom report where 12% are said to not have access to superfast broadband, which is at odds with their declaration of 89% superfast coverage,"

Dear oh dear. Of course it is not at odds.

There are likely to be some places where coverage is nil, and some 100%. The 89% is an average.

  • Fellwalker
  • 12 months ago

I think the reasoning of fellwalker is wrong, its more likely to be a rounding issue, since notice how rare it is for the report to quote coverage to a decimal place. 0.5% is some 145,000 premises.

If one was to do an average of say 400 local authorities you would get a very different figure to doing the maths as the whole UK. Why? Because not all local authorities have the same number of premises in them.

  • andrew
  • thinkbroadband staff
  • 12 months ago

@Fellwalker
The text you quote from the article is about places /in the report/ which use a 12% figure. @andrew talks about this being "at odds" simply because 89% + 12% does not total 100%, as you'd expect.

In your last paragraph, there are certainly places /in the UK/ where coverage is nil - especially if you define "place" sufficiently well.

But the 89% doesn't represent "an average of places": it actually represents the aggregate figures for the UK.

I don't think coverage is something where taking an average is a meaningful calculation. It seems likely to be an invalid use of maths.

  • WWWombat
  • 12 months ago

For example...
Assume London, 4 million premises, is 100% covered.
Assume Rutland, 15,000 premises, is 0% covered.

Average:
The average of two figures (0% and 100%) is 50%.

Aggregate:
Covered = 4,000,000 out of 4,015,000 = 99.6%
Not covered = 15,000 out of 4,015,000 = 0.4%

Big difference between the result of averaging places, and aggregating them.

  • WWWombat
  • 12 months ago

@wwWombat That use of averages also illustrates, in the opposite fashion, why Ofcom’s use of mean average rather than median gives a misleading overstatement when you are lumping together different types of service

Suppose we have 2% getting 1gbps, 3% getting 100 mbps, 35% getting 50mbps, 50% getting 15mbps, and 10% getting 5mbps the mean average speed is 48.5mbps, a speed more than three times the speed the majority of the population get.

  • gerarda
  • 12 months ago

Post a comment

Login Register