Broadband News

Does good superfast coverage always mean good broadband speeds?

Just a few days now until its time to look at the monthly analysis of the broadband speeds for well over sixty service providers and with the quarter coming to an end too we will have a new set of results to add to our broadband speeds and coverage site.

To mark the addition of a new table that will allow people to compare all the UK councils in a single table we have looked at the relationship between broadband coverage levels and the speeds people are recording using our speed test.

Chart of superfast coverage versus Q1 2017 speed test results
Click image for larger version

Relationship between Superfast Coverage and Speed Test Results

The link between broadband speeds people are recording on average across the UK council areas is an important one for several reasons and most importantly with public money being spent to improve superfast coverage areas in many areas councils will be key to see an observable difference. Another big factor is to verify where the announced take-up figures for superfast broadband is at all reflected in what the public is recording in speed tests.

The graph for superfast shows a fairly set pattern of speeds increasing until you hit the 90% mark where local authorities become more more spread out in terms of what we saw from our crowd based observations in Q1 2017. To see if this was due to the influence of ultrafast broadband coverage we have produced the same plot but ordered by ultrafast coverage on the x-axis.

Chart of ultrafast coverage versus Q1 2017 speed test results
Click image for larger version

Relationship between Ultrafast Coverage and Speed Test Results

The ultrafast plot does appear to show that the median download speed for a local council area is linked to the level of ultrafast coverage.

10 fastest local council areas in Q1 2017
Click image for larger version

Top 10 Council Areas for Median Download Speed

10 slowest local council areas in Q1 2017
Click image for larger version

Worst 10 Council Areas for Median Download Speed

The difference between the mean and median measures particularly for the slowest council areas needs a mention and this is due to the fact that with the take-up levels of superfast broadband those who have upgraded to a superfast (or ultrafast) option are skewing the results and this can also be seen the quartile (25%) figures. Once the Q2 2017 results are available we will be switching to displaying the 20th and 80th percentile figures for both provider speeds and speeds in the different parts of the UK. This is due to the expected changes in how broadband speeds will be advertised.

The Shetland Islands who have the slowest median download speeds only started to see take-up for FTTC/VDSL2 based services take off in Q1/Q2 2015 and as take-up increases the mean is likely to increase substantionally since in Q1 those with ADSL recorded a median of 4.1 Mbps compared to those on FTTC seeing 29.2 Mbps, eventually as take-up and coverage of FTTC increases substantion the median will also rise.

It is important that campaigners, lobbyists and policy makers understand that observed speeds will always lag behind the availability of faster broadband, this can be minimised to some extent by good local promotion of the better options that are available to people, but due to the nature of broadband contracts there will always be some lag. A key part is how the speed data is trending over time and our tracking goes all the way back to 2009, well ahead of the BDUK process delivering anything and also before the commercial superfast rollouts had delivered anything significant. 

Comments

There's some interesting analysis in here Andrew. Thanks.

Now that, nationally, the median broadband line is right on the cusp of being NGA, it makes me realise that the same will be (kinda) true for individual councils too. Those with a bit more NGA are likely to find their median speed being an NGA line, those with less NGA will find their median speed being an ADSL line.

However, as we're talking about speedtest results, we're talking about actual speeds, not available speeds. The median therefore comes from takeup.
...

  • WWWombat
  • 12 months ago

So I'd have thought that plotting councils and their median speed would actually show a relationship with the percentage of NGA takeup in that area.

Is "NGA takeup in a council area" something you have figures for?

  • WWWombat
  • 12 months ago

Working on something around take-up estimates, note the word estimate.

Along with lots of other changes to what data is available for those that want delve into things.

  • andrew
  • thinkbroadband staff
  • 12 months ago

Why use Council areas? Is that how the state provides its data (perhaps because of the confusion)? Is it a chalk and cheese comparison, or are the areas chosen purely incidental and irrelevant to the main purpose of the exercise just to show changes over time?? I'm thinking Parliamentary constituencies are reviewed more often, and specifically are supposed to reflect local characteristics and identity and aim for equal populations. Council areas are just history, 1972, or even 1933 or the nineteenth century, since the latter so-called review was just bunging smaller Councils together.

  • mervl
  • 12 months ago

Hi Andrew.
Would it be possible to number the list the same as the MP results as I record and check their position in Surrey at the 24 meg position % this gives a good indication where work is taking place and the average speeds are changing.

  • Blackmamba
  • 12 months ago

@Blackmamba I see no real point in the ranking anyway, since even if every LA was in the 99% region, you'd still have position 1 and position 650

We do change maps periodically, so those will tell you where the biggest coverage changes are happening.

  • andrew
  • thinkbroadband staff
  • 12 months ago

Why council areas?

Well we don't do just council areas...we also do constituency, and postcode groups e.g. GU8 and GU

Council areas are good as they generally line up with the various BDUK projects which are administered by Local Authorities.

Scroll down page at http://labs.thinkbroadband.com/local/ for constituency table
Councils are at http://labs.thinkbroadband.com/local/councils

Postcode letter groups at http://labs.thinkbroadband.com/local/map-tech#6/51.414/-0.641/

  • andrew
  • thinkbroadband staff
  • 12 months ago

Quartile figures are now on the stats pages e.g. http://labs.thinkbroadband.com/local/england you can view quartile, mean and median speeds both download and upload, and for the various technologies we can detect.

Once Q2 results are in it will change to 20th and 80th to line up with proposed advertising rule changes.

  • andrew
  • thinkbroadband staff
  • 12 months ago

Hi Andrews staff.
Thanks for the above remarks and your actions I have added the page position (25) to the 11 councils so when more post codes are added with fibre and G/Fast both results should change across Surrey. This will also show as the money from the OMR is being spent.

  • Blackmamba
  • 12 months ago

Post a comment

Login Register