Broadband News

Overall UK broadband speeds up 24% in the last year

Ofcom has published its Home Performance Analysis for November 2018 which states the average download speed was 54.2 Mbps and the upload speed was 7.2 Mbps. The download speeds rising by 18% compared to the last analysis a year ago.

Looking at our own data from the extensive speed testing that the public do using our speed tester and partners that use our technology the basic average mean speed rose to 34.6 Mbps in Q1 2019 a rise of 14% in the year. If you use a clean figure which should exclude poorly performing connections and some wireless testing the mean rises to 39.5 Mbps and has increased by some 24% since Q1 2018.

Another way of looking at the speeds from the consumer testing is to look at the speeds the top 20% are getting which is 50.2 Mbps down and 11.3 Mbps upload. This seems an odd choice but we know from looking at the devices used around half of devices are tablets or mobile phones in the testing, we can exclude them but you still have a lot of people browsing on laptops over their Wi-Fi, so looking at the top 20% speed is more often than not going to be those who have everything connected properly and thus more closely mirror the optimal solution that the SamKnows solution creates i.e. ethernet connected device to a well known connection and the results of each of the 4000+ boxes across the UK are then modelled to produce the final figures.

So after a few paragraphs what we can agree with Ofcom on is that UK average speeds are increasing and the reasons are simple more people are signing up for faster services, or in some cases people are being given free upgrades. This is nothing new at all hence the climbing mean average speed shown on our broadband information site which now shows the average overall speeds going back some ten years along with how the different core technologies are performing.

Getting more people to sign up for the faster services is something to be encouraged. This includes encouraging those who while they may not get superfast speeds today due to their distance from a VDSL2 cabinet but can often still get substationally higher speeds compared to their old ADSL. Ofcom is also keen to point out that if you are paying standard out of contract pricing you may actually end up paying less switching to a VDSL2 service.

The challenge of getting the public to sign up to the best service is highlighted by the gulf between the theoretical maximum for the UK which is 208 Mbps and the observed 34.6 Mbps and this keeps creeping up as more full fibre and cable services are rolled out. Areas such as Herefordshire which are towards the bottom of the superfast availability table have a much lower maximum mean of just 89 Mbps versus the observed mean of 25.5 Mbps.

The take-up of superfast services is improving hence the increases and Ofcom reports "In 2018, around two-thirds (66%) of UK home broadband connections were superfast products with an advertised speed of 30 Mbit/s or higher."  This statement does not mean two thirds are on actual connections with speeds of over 30 Mbps, but the speeds in the adverts are over 30 Mbps, so those with just 15 Mbps but buying an advertised 36 Mbps are in this figure. Looking at the technology splits we see from speed test visitors (remember there is no modelling applied to our figures) and our figure of 74.9% (combined FTTC, cable and FTTP figures) while higher than the official figure is likely explained the tendency to take your shiny new fast broadband connection for a test drive.

The takeaway figure that the industry, investors and analysts need to pay attention is that only 1% are on 300 Mbps and faster products (ultrafast using Ofcom definition) and this is when speeds of 300 Mbps or better are now available to over 50% of the UK population. Ofcom recently reported the 50% mark being passed, our ultrafast coverage figure is higher since we count 100 Mbps and faster as ultrafast, the Ofcom figures end up excluding areas where Virgin Media is only selling its 200 Mbps service (a choice they make to help manage peak time congestion). So while for some faster speeds are paramount we are in the situation where only 1% are choosing it when 50% have the option and this represents the marketing challenge of the next decade, price is a massive factor since the public will invariably choose the cheapest service that allows them to do what is their core use of broadband which for many today is can they watch a film in HD from a streaming service.

This wide choice of speeds and price points partly explains why the UK is so far down global comparison tables, this will change as the new market entrant products such as Vodafone Gigafast with its 100 Mbps slowest service and others with single speed Gigabit products gain market share and in many other countries that change has already happened.

We can blame BT Group for the UK being so average, but the Digital Britain report of 2009 was the kickstarter document that lead to the focus since 2012 of getting reasonable superfast speeds to as many places as possible. Decisions by BT to focus on VDSL2 in its commercial roll-outs and shrink its FTTP ambitions have only been reversed in the last year, when unsurprisingly less resources are needed to meet contractual obligations for the local authority BDUK contracts. The other big factor was the rise of TalkTalk in 2006 and the LLU market place which seems to have meant investors are only in the last year or two looking to gamble on the next step which is full fibre - if LLU had not happened broadband in the UK would be more expensive but we would likely have seen more infrastructure competition.

Excuse the lack of pretty graphs, could easily have added half a dozen but all the important stuff is on our stats site and with its breakdowns of the trends over time for local authorities looking there will show information much more relevant to the individual.

The key takeaways are:

  • Encouraging take-up of better broadband services is important
  • Ofcom average speeds may not always agree numerically with our speed test figures, but the trends are remarkably similar, thus showing that things really are changing even if for some several hundred thousands of premises it seems like nothing has changed for a decade.
  • We are entering an uncharted era with some parts of the UK having extensive full fibre availability (extensive defined as 75% and higher) or reaching that very soon.
  • Full fibre operators if you want people to feel the real impact you need to be providing top notch wireless capabilities.


There seems to be a stray 'ofcom' in para 4.
Para 5: 'substationally' - also the entire para seems to be one sentence and have no commas.
Sorry to pick nits...

  • Diggory
  • 10 days ago

"even if for some several hundred thousands of premises it seems like nothing has changed for a decade."
Nothing has changed for a decade, only instead of being half average speed, we're now a twentieth of average speed.

  • brianhe
  • 10 days ago

Don't mind my proof readers pointing out that :-) We are all in it together.

  • andrew
  • thinkbroadband staff
  • 10 days ago

"even if for some several hundred thousands of premises it seems like nothing has changed for a decade."

Indeed, there is nothing whatsoever about the qualification, "seems" ; for hundreds of thousands of properties, this is a hard reality, not a subjective statement. Until legislation is used to drive a fair and equitable distribution of broadband provision, the gap that brianhe mentions above will just increase.

How about a service charge that is proportional, across the UK, to the actual speed seen on each premise?

  • p6resthome
  • 10 days ago

@p6resthome "How about a service charge that is proportional, across the UK, to the actual speed seen on each premise?" or how about a charge that reflects the cost of installing, providing and maintaining that service? I suspect that the cost of installing and providing a "fast" urban service is in many cases considerably less than that of a slow rural service.

  • MCM999
  • 10 days ago

I used the word seems, because in absolute terms the upgrades from ADSL to ADSL2+ will have uplifted upload speeds for almost all and download speeds for many but not into the superfast era. Not much but enough to mean some things have changed.

There is also a growing number of alt net particularly fixed wireless that say they can serve many of these of these underserved areas

  • andrew
  • thinkbroadband staff
  • 9 days ago

It might be worth some comment on the VDSL degredation effect with more take up. I'm only one example, but with no changes I've gone from an approx 60mbps sync to a 40mbps sync in 2 years.

Seemingly as neighbours have taken up promotions from Satellite and moved from Cable or even upgraded from ADSL to VDSL - and increased crosstalk.

  • jchamier
  • 9 days ago

While some individuals are seeing crosstalk effects those with vectoring should not be.

The reason why not commenting is that the overall data does not say VDSL2 is slowing down en-masse

ECI and Huawei areas are holding steady speed wise, with a trend upwards - likely driven by vectoring, G.INP and lower target margin effects.

Our projects of speeds for VDSL2 have always taken a pretty pessimistic view of cross talk too

  • andrew
  • thinkbroadband staff
  • 9 days ago

Post a comment

Login Register