Ofcom publishes its Connected Nation Report for 2017

The Ofcom Connection Nations report[1] is seen by many as the annual de facto status report for whats happening in the UK and looks at both the services that are available to order and the speeds that people are getting, something that has often been confused in last 12 months since the last report from Ofcom. The headlines from the report are:

  • UK superfast broadband coverage in May 2017 was at 92% for the over 24 Mbps definition
  • UK superfast broadband coverage in May 2017 was at 91% for 30 Mbps and faster, a 2% increase compared to 12 months ago
  • Average download speed was 44 Mbps download and 6 Mbps upload compared to 37 Mbps and 4 Mbps in 2016
  • Scotland saw superfast coverage increase 3% to 87%
  • Wales saw superfast coverage increase 4% to 89%

Of course as regular readers will know, we also are in the game of tracking broadband coverage levels and have been since 2012–our figure for the UK on 30th April 2017 was 92.8%. What is missing from the Ofcom data is any indication of decimal places and whether their figures apply the usual rounding or are taking floor values.

Similarly there’s no confirmation on whether speeds are median or mean averages, so it entirely possible our figures were within a fraction of one percent, or a bit further apart. Our up-to-date statistics on our Local Broadband Information[2] site are usually as low as a few days behind when we start seeing users on new faster connections. There is a recognition of the work we have been doing in terms of tracking coverage levels in the UK in the report but there is potential for people to be mislead in what we do.

We had been expecting a mention, but more around that while two different methodologies are used that the figures are within a reasonable range.

2.11 This figure is lower than some numbers reported elsewhere, such as by thinkbroadband.com, whose figures are cited by Government to track superfast broadband rollout. Unlike our estimate, which is based on analysis of speeds available to every residential and small business property in the UK, other such reports are extrapolated from consumer line speed tests for a small sample of lines. Ofcom Connected Nations Report page 12 section 2.11

We expected to be within 0.5% of the Ofcom data analysis, but this paragraph from the report gives a false picture of how we arrive at our coverage data and therefore it is time to make it clear what we do.

  1. Government has access to data from operators (possibly the same data as Ofcom) and is able to track the coverage levels and this is an important part of the tracking the BDUK costs and also why DCMS is confident on reaching the 95% superfast target by the end of this year.
  2. This means where they are quoting our data, and similarly for local authorities we are referenced as an independent source and therefore why we will state that while providers may have said they have delivered the coverage by the end of December, we want a few weeks to independently verify that this is the case.
  3. The Ofcom analysis is done at the premises-level and our analysis is presented at the postcode level, with postcodes being flagged as superfast or not based on the majority of premises (we don’t ask for full postal addresses for privacy reasons).

    In previous years this gave wider error margins particularly in city centre areas due to Exchange Only lines but the number EO VDSL2 cabinets appearing and network rearrangement that has happened in 2017 means this issue is less of a problem in statistical accuracy terms than in 2016.

  4. We do not directly extrapolate coverage data from speed test results; these form part of our methodology in identifying superfast coverage however the sample size concerned raised by Ofcom are invalid.
  5. Our speed test is not a line speed test, but it is a series of measurements on the observed download and upload throughput for a period of time. We are able to identify the technology in use for a speed test with a good degree of confidence.
  6. Our sample of lines could not be described as small; we believe that the speed test data we hold is the largest accurate data set for the UK in terms of broadband performance by area. Although there are some speedtest-focussed sites which may have more raw tests, these cannot correlate such data down to our level of detail.

    Based on our UK-specific knowledge, we are able to provide a much more informed insight into performance for consumers and are very visible when it comes to asking questions about a specific test.

  7. The year 2017 has seen a level of unpredecented work in verifying our model and ensuring that the many variations that do exist are catered for, rather than the model being a simple extrapolation system.
  8. While many media outlets will take up stories of consumers unable to order FTTC/VDSL2 services, we have a steady stream of contact with providers to solve ‘computer says no’ issues and these are most common for full fibre. The end result of this is that we know that providers’ own data systems are never totally accurate, and in fact no large data system will be error free. As we publish our UK postcode level data on our local broadband site[3], we do see more consumer enquiries and we do investigate these to ensure our data continues to remain the most live.

We do use a very different methodology to Ofcom and DCMS since we do not have access to the same data set which they originate their analysis from (which we’d be happy to ingest if the industry want to provide this), however we know that even that data set is not entirely accurate, as we feed back issues to providers.

The advantage to not using supplier-provided data (some smaller operators do supply data, but we still do periodic checks on these) is that there is freedom to use the data and share different types of analysis. Of course given the political importance of the 95% superfast broadband target, there is a statement from the Minister for Digital in response to the Ofcom report.

We will have taken superfast broadband to 95 per cent of the UK by the end of the year but we are not complacent and there is more work to be done. We are making sure that by 2020, every home and business in the UK will have access to reliable, high speed broadband.

There is also a clear need for rapid improvement on mobile coverage. We’ve recently removed outdated restrictions, giving mobile operators more freedom to improve their networks including hard-to-reach rural areas. But industry need to play their part too through continued investment and improvement in their networks, making sure that customers are not paying for services they don’t receive.

Minister for Digital Matt Hancock

The Ofcom Connected Nation report is of course much more than just a single factsheet, but given how we have seen those specific facts used and extrapolated in the last couple of years we feel we had to make it clear where we stand on their superfast statistics claims.


Post a comment

Login Register[4][5]


  1. ^ Ofcom Connection Nations report (www.ofcom.org.uk)
  2. ^ UK Broadband Information (labs.thinkbroadband.com)
  3. ^ UK Broadband Local Information (labs.thinkbroadband.com)
  4. ^ Login (www.thinkbroadband.com)
  5. ^ Register (www.thinkbroadband.com)

Leave a Reply

Your email address will not be published. Required fields are marked *