BIG claims that 6.7 million broadband connections fail USO test
Bad broadband is a very easy headline for an article and even better if you can talk about millions of premises affected and then extrapolate this to suggest that each of these is owed £25 compensation as a result, meaning millions will read an article thinking they have a windfall on the way.
Broadbad 2.0 from the Rt Hon Grant Shapps MP and BIG has been published and attempts to make a case for automatic compensation for broadband customers.
If Mr Shapps is reading this we are more than happy to meet and explain what our data can tell him about broadband in the UK and what the official Ofcom and Government figures also mean, we have 17 years of experience in UK broadband and deciphering the PR from the reality.
Before we delve into the report a quick summary of what is actually happening currently, Ofcom is already looking at automatic compensation for total broadband outages, the Universal Service Obligation is progressing and no-one has been given the role or stepped up with a plan that is acceptable to Government or Ofcom yet on delivering the USO.
Now we will dissect the report and its key points made in the executive summary, since without any challenge it would be very easy for people to accept them as fact and get angry with their provider when the reality is very different.
Text in bold/italic is a direct copy from broadbad 2.0 report:
The current available data recording broadband speeds needs to clearly distinguish between the take-up and availability of connections above 10 megabits (Mb/s). BIG has found that as many as 6.7 million UK broadband connections may not receive speeds above the proposed minimum standard of 10 Mb/s. However, this data, recorded by industry regulator Ofcom, fails to distinguish between those customers that have actively chosen not to take up superfast broadband, and those superfast broadband customers that do not receive the speeds they pay for. It is therefore almost impossible to determine from Ofcom’s data how many customers do not receive the speeds that they pay for. This report consequently calls for greater specificity and transparency in data recording broadband speeds.
The data used to arrive at the 6.7 million is the connection speed data from the Ofcom Connected Nations reports, the latest of which was published in December 2016 (the data was released in early 2017) and is based on data collected from broadband operators in the first half of 2016, and thus is over 12 months old now.
6.7 million connections is a lot higher than the number of premises that are under 10 Mbps based on the availability data Ofcom publishes, and over six times higher than what we are tracking currently where 904,000 premises are under the broadband USO (3.15% of UK premises, when looking at just fixed line services and has dropped 0.02% in the last week, fixed wireless services which easily beat 10 Mbps will drop this figure even further). This difference is explained by the reality that upgrades to faster services are not automatic, hence you will see a big difference between coverage and take-up figures - something the Broadbad 2.0 report fails to fully understand.
The report calls on Ofcom to improve its data collection, but to really work out who would 'deserve compensation' it would require actual input from every broadband customer to understand why they have not ordered for example a Virgin Media cable service and prefer to remain on an ADSL2+ service at under 10 Mbps. Additionally speeds vary due to the shared nature of consumer broadband, and thus data collected one day is out of date the next, plus there are situations where people are getting slow speeds due to something under their own control e.g. slow Wi-Fi devices or presence of poor telephone extension wiring.
In theory we could do some analysis and look out how many of the speed tests we see are in sub 10 Mbps areas and how many are in areas with superfast broadband, but this would not tell us why people in the superfast areas are only getting below 10 Mbps, i.e. to fully understand you need to carefully question people and with around 23 million broadband connections in the UK this would be a major undertaking.
Broadband providers are failing to deliver quality customer services.
This BIG investigation into broadband customer services has found that the internet providers signed up to Ofcom’s Voluntary Codes of Practice are unable to provide any figures for the average amount of compensation that they pay to their customers when something goes wrong with their service.
The current Ofcom Broadband Speeds Code of Practice does not specify compensation but rather it operates so that consumers will have the ability to exact contracts early, or if a provider has a cheaper product that more closely matches the speeds the customer is receiving they should be allowed to downgrade.
Compensation for broadband outages is paid automatically to broadband providers when its the upstream provider at fault, but the procedures for consumers claiming this varies greatly and Ofcom is already consulting on changes in this area and the compensation due will vary based on the length of outage, contract the consumer has with the retailer and the underlying cause.
Compensation under the current system varies greatly, and the best providers do things like send out 4G mobile devices to keep people online.
The danger if compensation culture was to be the norm is that people would be trying their best to slow down their connection just enough to claim compensation and turn their broadband into a cash generation system, and broadband providers would simply end up raising prices across the board. At the end of the day we are sure the public would rather providers be fixing their broadband than getting some compensation rather than spawn another set of spam telephone teams offering to help you claim the compensation that is 'rightfully yours'.
Less than half of all UK broadband connections reach speeds above the ‘superfast’ threshold of 24 Mb/s. BIG analysis of nationwide broadband download speed data recorded by Ofcom in 2015 and 2016 reveals that 40.8% of all recorded UK broadband connections reach speeds above 24 Mb/s, the threshold for superfast broadband. Moreover, the 20 worst constituencies for slow broadband download speeds have not changed significantly over the calendar year 2015- 2016.
This is the biggest flaw in the broadbad 2.0 report, or if this was the United States we would be saying it was fake news. The report is conflating take-up and availability to some extent still, though the wording is better than a copy of the report since a week earlier. The full report is pretty clear, but given the time pressures at which the non tech press work under, it is likely that we will see take-up and availability being mixed up.
The main problem here is that they are using data from 2015 and 2016 and making claims of no significant changes, so using our coverage and speed tracking we can see Carmerthen East and Dinefwr has seen superfast coverage jump from 37.9% in July 2015 to 56.8% in July 2016 and is currently running at 65.1% with full fibre coverage of 5.8%, observed mean broadband download speed in Q2/2015 was 10.4 Mbps, Q2/2016 11.6 Mbps and 14.3 Mbps in Q2/2017, so speeds are improving and higher take-up rather than compensation will push things higher. In Q2/2017 as the availability of full fibre in the constituency has risen we are now reporting the average download and upload speeds in the constituency too. The latest financial report from Openreach showed that in the last quarter across the UK they added 437,000 new fibre connections so the scale of the change is easy to spot and thus using data a year or two old is never going to tally with what operators say they have delivered.
In Q2/2017 of those running a speed test on thinkbroadband or any of the sites using our testing technology we saw 37.8% of testers using an ADSL2+ connection, and the same on VDSL2, then 22.6% on a cable broadband service and 1.5% on full fibre. The proportion of VDSL2 and FTTH tests is rising and our https://labs.thinkbroadband.com/local/ site gives a breakdown for each council/constituency area of the different ratios actually seen alongside what we know about the availability. Most importantly the site shows the 20/80th percentile in addition to mean and median speeds so people can judge for themselves what people are using in an area and even what the speeds are for FTTC versus FTTP versus ADSL2+.
BIG therefore calls on Ofcom to urgently consider the inclusion of fixed broadband speeds in a new automatic compensation scheme, and on the government to progress secondary legislation to the Digital Economy Act that sets out the terms for a broadband USO.
We are not sure what BIG means when they talk about 'fixed broadband speeds' at several places in the report, it might be referring to fixed line broadband technologies i.e. ADSL2+/VDSL2/G.fast/FTTP/DOCSIS rather than mobile, or it could be talking about fixed connection speed services only such as DOCSIS and FTTP with FTTP generally not dealt with by Ofcom due to its low availability, i.e. Ofcom regulates for the most widely available services, though excluding ADSL2+ and VDSL2 form automatic compensation would be odd.
A report calling on the Broadband USO to be defined is very odd as that is exactly what is happening. The calls for the Broadband Speeds Code of Practice to made mandatory is something that Ofcom is actually considering but it wants to walk the voluntary path for its next wave of changes first, the broadband market is so diverse that mandatory legislation can actually cause as many problems as it solves. The broadband market is at a nexus point and too much regulation may stiffle full fibre roll-outs which are just starting to scale up to deliver some serious volumes at last.
Improving data collection is interesting, Ofcom already gets a raft of connection speed data from broadband operators for fixed line services, but connection speed tells you nothing about the actual experience, hence why Virgin Media was recently on Watchdog for broadband performance failings even though they are fixed speed connection service. As part of the two months of work ahead of the actual broadcast we made available a map that highlights the variation in quality of service for the main UK broadband technologies (download, upload, quality and latency data is featured). So short of making speed testing of connections mandatory it is hard to see what else can be done without a lot more time and people involved, which then means higher costs.
The harsh reality of consumer broadband and why it is so cheap per month is that it relies on connectivity methods where bandwidth is shared, i.e. no consumer broadband provider can provide all its customers with their maximum throughput speeds 24/7 and while the amount of capacity budgeted is increasing at many providers we are still looking at a budget of 1 to 2 Mbps per customer. If legislation is to be enacted to guarantee every broadband customer in the UK was able to get at least 10 Mbps 24/7 then prices would rise substantially.
For those who have got this far through the article and also have very slow broadband and no option forof anything faster, we are well aware of your plight and hope that the USO will deliver meaningful upgrades, and while talk of compensation may please some it does nothing improve your broadband connection when what you want is providers to offer higher speeds.
A more productive report would have looked at areas of those constituencies where full fibre broadband is available and research what differences consumers and business users are finding in terms of reliability, speeds and general productivity and could have been used to help drive the case for much more investment in full fibre broadband by both BDUK gap funded projects and the commercial operators.
A final take-away, automatic compensation for slow broadband is likely to mean that providers will be much more picky on who they sell broadband too, and thus people in slower areas are likely to find their provider choice even more limited that it is today, and it also could jeopardise the broadband USO because it will be seen as a poisoned chalice that no provider wants to take on. A very real alternative would be to sell services such that the minimum guaranteed speed was so long that the compensation would be useless.