Broadband News

Which? leaves us confused with its latest broadband news

This month is feeling like we are working as the guardians of broadband statistics with a broad range of data being thrown around and bold claims abounding - alas boring headlines tend to vanish but bold claims of people being massively undersold will be sure to gain newspaper and wider coverage.

Which? appears to have reprocessed the results from consumer broadband speed tests collected in Jan to March 2017 which it originally published in June, but confusingly the numbers appear to have changed, so that in todays coverage West Dunbartonshire is described as getting an average speed of 41 Mbps (Mega bits per second) but back in the original June reports the median download speed was said to be 20.7 Mbps, top quartile 43.6 Mbps and a mean of 31.1 Mbps.

One possible reason for the difference is that the new push from the consumer group is trying to make the point that if selling broadband at speeds that only 10% of customers are expected to receive that millions will clearly not get that speed. Their analysis has said 'the overall speeds recorded in broadband speed tests were only 58% of the speed that users were expecting to get' which means providers overall are exceeding the current 10% rule, since the result should be that only 10% of users are getting speeds above the advertised figures.

The latest piece of analysis from Which?, as part of its Fix Bad Broadband campaign, shows the overall speeds recorded in broadband speed tests were only 58% of the speed that users were expecting to get.

The results show that the faster the expected speed, the bigger the expectation gap between what consumers thought they should receive and actual speed recorded in the tests. For example, consumers expecting speeds in excess of 30Mbps (between 30Mbps and 500Mbps) were on average only able to get 54% of the speed they were expecting.

Users expecting internet speeds of between 10 to 30Mbps received 89% of what they expected. And those expecting to get 10Mbps or lower were able to exceed their expected speed by 38%.

Which? on broadband speeds

Revisiting West Dunbartonshire Which? has today said the average expected speed in the area was 83.7 Mbps and when you look at our data for the area which goes all the way back to 2009 its clear that this high figure is because Virgin Media cable broadband is so widely available in the area and on our speed test results for Q2 2017 we saw twice as much Virgin Media testing locations as we did FTTC/VDSL2 in that area and the period of the Which? results covers a period when Virgin Media was experiencing more issues than normal.

Previous Ofcom research has previously found that even when conducting face to face interviews with broadband consumers that people are often not aware of the speed of the package they have purchased, so asking consumers to select from a drop down list what is the current broadband speed they pay for is very likely to produce some very interesting results with some people picking the advertised speed, others picking the point of sale estimate and some picking a random answer. On the mystery of why the speeds for West Dunbartonshire and others are different it is probably because some people were picking the 'Other Not Listed' option and thus their speed test was excluded from the latest analysis.

The analysis still pre-dates the fix bad broadband campaign and if Which? do eventually release any analysis from that campaign they are asking people to pick from a different range of speeds so data will never align and the list of providers is far from complete.

We regularly explore various questions to do research for ourselves and have found the public are frequently worse at picking their broadband service type and advertised speed compared to what we have spent lots of time automating e.g. when we've done this to samples of visitors we have seen things like Virgin Media customers claim to be on an up to 38 Mbps service - a speed they've never sold or people who we can say with certaintity are on a JANET connection claiming to be using BT.

We have reached out to Which? on twitter previously over concerns with the accuracy of the broadband speed test which is reporting unusually high speeds for some, there is no sign that their test supplier had improved the test and this latest extrapolation from what is likely flawed data looks much more like headline chasing rather than helping to inform decision makers. Therefore this latest PR push is not going to help produce a useful change in how broadband is advertised, if anything providers as a result of coverage like this are likely to talk even less about broadband speeds and switch to the old fall back of lifestyle marketing, i.e. style over substance.

So what should the readers of this article do? Well the key message is when signing up for broadband it is often best to avoid actual sales staff as sales patter can distract you and has a greater potential to mislead than signing up online where you can take your time and digest things like the personalised speed estimate that providers signed up to the Ofcom Broadband Speeds Code of Practice should supply at the point of sale, taking a screenshot of this estimate is worthwhile just in case a provider disputes it at a later date. If you think you are being undersold, running a number of broadband speed tests and making sure you run these at various times of the day is a good step and connect via an Ethernet cable and ensure nothing else is using the connection when testing, the broadband providers support will invariably run through a raft of basic diagnostics to try and ensure your slow speeds are not basic things like something plugged into the phone line without a microfilter, or you have your broadband modem on the end of a long extension which is killing your potential broadband speed.


These poorly conducted tests just leave a get out clause for the ISP`s as they can easily dispute the findings and muddy the waters even more.

  • pipcoo
  • 12 days ago

which is more or less pointless to use as its payed only customers who can use it (like recommending giffgaff when data speeds are poor typically due to inadequate bought capacity and can get better deals elsewhere for same price)

people that are doing these have likely forgotten what the ISP told them what there estimated speed was at point of sale (excluding virgin media or actual real FTTP as speed is profile based not distance)

  • leexgx
  • 12 days ago

Have had the unfortunate experience of using the Which? speedtester. It took weeks to work for me then, when it did, provided info that was at best 'pie in the sky' e.g. that a neighbour could get 76Mbps when in fact my hamlet is not even fibre enabled! As for the speed it suggested I was getting.......I wish!

  • BBSlowcoach
  • 8 days ago

Certainly something odd about Which? BB speed test results. I tried it twice and got:
1st test - Response Time: 301ms Download Speed: 0.5Mbps. Upload Speed: 11.3Mbps
2nd test - Response Time: 285ms Download Speed: 0.5Mbps. Upload Speed: 4.2Mbps

My router is synchronized at (Down/Up) [kbps/kbps]: 608 / 440.
The upload speeds are bizarre. You can see from the sync. speed that 0.4Mbps would have been nearer the mark.

  • willbard
  • 5 days ago

Post a comment

Login Register