We’ve had it ‘up to’ here with broadband ads

The disparity between different customers’ speeds with the same ASDL provider is mostly down to distance from the BT exchange. This is an inherent problem with ADSL which is outside ISPs’ control, so I don’t think it’s fair to blame them. Fibre to the cabinet (FTTC) eliminates this disparity, because the distance from the cabinet to the premises tends to be more uniform and much smaller. However, it’s high time that we move away from archaically carrying the internet over copper telephone lines and instead carry telephone lines over the internet using fibre to the premises (FTTP).

Broadband providers should also advertise their upload speeds as prominently as their download speeds. Some upload speeds are shockingly low, often just a tiny fraction of the download speed. This means that if you want to send a large video, for example by uploading it to YouTube or sending it to a friend via Skype, then it can take forever. However, this is what the A in ADSL means – asymmetric; the download and upload speeds are not the same.

The figure of 10% also has another completely different relevance, which is not well known by consumers. Your maximum bandwidth will always be around 10% less than the speed of your connection because of control and framing data. Think of it a bit like an envelope in the postal system. Although the contents of a thick letter might weigh 90g, the total weight including the envelope might be 100g. For example, if your ADSL line sync speed (viewable in your router’s web interface) is 10Mbps, then you will never get more than around 9Mbps of bandwidth. Even my gigabit broadband connection at home can’t get more than around 900Mbps on a really good day.

PeterM says:

“Fibre to the cabinet (FTTC) eliminates this disparity” – sorry, that’s not correct. There are still issues concerning distance, and in part what cable is used. I’m in a semi in a cul-de-sac, served by aluminium (not copper) cabling that goes “around the houses” (actually 5 blocks of flats, 2 low, 3 very high, so around 300+ households at a guess) to a cabinet around 600 meters away.

Speed estimate on FTTC for my home is 33.5 Mbps, but the takeaway across the road behind my house (and their fibre cabinet is 30 feet from my back window) has an estimate of 67 Mbps. It’s down to both distance and cable type (and many, many problems, according to some of the Openreach engineers I’ve spoken to).

You can also find some locations in the UK which will not get even 30 Mbps, let alone 60-75 Mbps, where their FTTC is only offering 15 Mbps. Down to location, cabling, etc.

I agree with all your other points, but there are quite a variety of speeds available even on FTTC (and I agree, if it was the other type, Fibre To The Premises, the speed would not be susceptible to the vagaries of the cabling or interference like electric fences, Christmas lights, etc), and felt that people should be aware there is still no “guarantee” of the highest (75 Mbps and above) speeds for ALL….

User 66219 says:

4 January 2015

This comment was removed at the request of the user

wavechange says:

I am not happy with advertising broadband speeds that only 10% can achieve. Let us put the customer first and advertise speeds that 90% can achieve. Those who get a better service will be pleased rather than the majority feeling short-changed.

When I first signed up for broadband I contacted my ISP that my speed was only 0.8Mbps, 10% of the advertised 8Mbps and was actually told that I should be happy with this. When I made it very clear that I was not happy, something was done at their end and the speed improved considerably, though it was still less than half the claimed speed.

My ISP no longer advertises ‘up to 24 Mbps’. It gives a typical speed based on postcode and a maximum and minimum speed to expect. That seems much fairer.

NFH says:

I don’t see how ADSL ISPs can advertise a useful average speed, because it’s so dependent on the distance from the exchange. The only meaningful speed to advertise is the maximum theoretical speed, which will be reduced according to the distance from the BT exchange. Most ISPs provide a more accurate speed prediction based on a prospective customer’s phone number and postcode. Generic headline prices should mention only the maximum theoretical sync speed of the ADSL technology used.

The problem I do have is with ISPs that do not invest in sufficient infrastructure to support the number of customers they have. This results in contention with too many customers using limited network infrastructure at the same time. To avoid this, I suggest avoiding the large well-known ISPs that compete mainly on price to consumers with limited technical knowledge and instead go for a niche ISP that is favoured by very technical consumers.

wavechange says:

“The only meaningful speed to advertise is the maximum theoretical speed, …”

I totally disagree. This is worthless information to customers and the reason there have been so many complaints. It’s the reason why we are having this Conversation.

After years of lies, we need ISPs to deliver what they promise, or more. There is nothing wrong with giving customers at least what they expect. It might command a bit of respect.

NFH says:

Only the maximum theoretical speed of the technology allows consumers to make an informed choice between each technology. For example, ISP1 uses standard ADSL which supports a maximum of 8Mbps and ISP2 uses ADSL2 which supports a maximum of 24Mbps. In a particular consumer’s case, both speeds might be reduced by 30% because of the consumer’s distance from the BT exchange, but this 30% would be the same across all ADSL ISPs. This approach applies only to comparing between different ADSL ISPs, which is unfortunately still the most prevalent broadband technology used in the UK.

When comparing between different technologies, e.g. ADSL, FTTC, FTTP, WIMAX, satellite etc, then yes, the average speed does make a difference.

It’s important that consumers understand that broadband speed is affected by:
1. Technology used
2. Distance from the exchange (in the case of ADSL)
3. Contention ratio used by the ISP

Given that 2 is outside ISPs’ control but affects the speed so greatly, it’s not meaningful to use it to compare different ADSL ISPs.

Publishing an average speed oversimplifies something that can’t be simplified.

wavechange says:

The ISP needs to take into account the distance and other factors in telling us what speed we can expect.

As I have said, it does not matter if they underestimate the speed but it is dishonest to significantly overestimate it.

My ISP tells me that the average speed for my postcode area is 10Mbps and that the speed may vary between 8 and 12Mbps. My speed is at the lower end of this range, but elsewhere in the street, speeds are higher. It is honest and very much better than quoting ‘up to 24 Mbps’, which is what the ISP used to do.

It’s time to stop short-changing customers.

malcolm r says:

My ISP told me it would need a phone number to most accurately predict my likely broadband speed. Not a post code – does not give as accurate a result.

I understand what “up to” means, but I also want an estimate that predicts my particular property. Asking the right question seems to elicit the right information – nothing to do with honesty. But we probably need to be prompted to ask the right question in the first place.

wavechange says:

I accept that a phone number is better, particularly in rural areas where a postcode can cover a large area. I don’t care how an ISP estimates what speed they can provide as long as they can provide useful information.

I understand about distance, contention ratios, signal to noise ratio and some of the other factors involved, but what matters to each and every customer is what speed they can expect to achieve. We should not have to ask technical questions.

Lets have an honest estimate of what speed can be provided and in the interests of customer satisfaction it is better to underestimate that speed. It’s all about honesty.

malcolm r says:

“We should not have to ask technical questions.”. If a prospective ISP states that to give you your best estimate of broadband speed they require your phone number, that is hardly a technical question. If you then get nothing like the estimate, then you can criticise them. We need to play our part in this.

malcolm r says:

We keep returning to this topic. Before you sign up to a broadband contract you can get an estimate of what you are likely to get at your particular property. Post code is not accurate – telephone number is more useful.

As NFH points out there are physical factors that affect it. And an average is nearly as useless as a maximum – I want to know what I am actually likely to get.

What might be useful is if broadband providers pointed out the means by which you can get an estimate from them before taking out a contract.

Perhaps Which? could publish a helpful article on how to make the best choice of broadband, and stave off the rants that all these providers are dishonest charlatans.

wavechange says:

I was one of those who used to accuse my ISP form making a misleading claim about broadband speed of ‘up to 24Mbps’. I now congratulate them for no longer making this claim for their ADSL2 service and delivering a speed that is consistently within the expected range. All ISPs should be honest and if necessary underestimate the speeds we can expect to achieve. If it was a national service, I would be promoting it for honesty, albeit after a poor start.

I would like to have a cooling off period of 14 or 28 days if their ISP fails to deliver. The complication is that internal wiring and other local factors can seriously affect speed in some cases.

John Ward says:

We have Fibre To The Premises.We have paid for it as part of the price for buying a new house. I’m not sure I fully understand why BT and others now need to charge a premium for supplying much faster broadband when it seems to me it would to cost them more to slow it down! There is also a £30 ‘activation fee’ with BT which must be the cost of clicking a switch these days. Since we require none of the enhanced speed functionality we have chosen standard broadband and are content with the speed of just under 6Mbps, although this is lower than BT’s estimated broadband download speed range of 6.5-8.5 Mbps.

wavechange says:

BT can charge more for providing a faster service for those who need it. It seems fair to me, and a way of funding the roll out of fibre broadband.

I don’t understand why you should be asked to pay an activation fee if your new home was connected to fibre broadband from the start.

malcolm r says:

Perhaps it would be useful for Which? to ask BT. It may be both making a connection at their end and the cost of setting up a new account. I don’t know – let’s ask?

John Ward says:

Thanks Wavechange and Malcolm. I expect there is more to do than flick a switch to activate the BT Infinity service as something has to be installed [or included in the circuit] at the exchange end to enable the higher speed data transmission and they might also supply a new, more powerful router. In the overall scheme of things £30 is OK, but of course there is a considerable additional monthly charge [over & above the line rental] for the superior service and a major part of that is probably used to repay the capital investment in the the broadband roll-out which has indeed been colossal. However, there seems to be no ‘happy medium’ at a reasonable price – it’s either standard broadband, typically at up to 8Mbps, or superfast at anything from 38Mbps upwards and the higher speed services have other – possibly unwanted – things in the bundle and longer contracts. As Wavechange wrote recently on mobile phone coverage, it would be very useful if it were possible to take the service for a trial period to see whether or not it was worth the difference and actually delivered. As they always say on TV, “other operators rates and charges may differ”.

PeterM says:

Think you’ll find there’s a bit more to do than ‘flick a switch’ electronically. For a broadband service to work, whether FTTC or ADSL, needs the data and voice connections to be linked. The location of that link depends on what one gets (*). For ADSL the links are made at the exchange. For FTTC, the link is done at the street cabinet, and done by an engineer, so the human touch is still required.

Setup used to cost 50 quid (usually absorbed for new customers on 12, 18, 24 month contracts) but OFCOM forced Openreach to charge properly (half for setting up, half if one ‘ceased’ the internet service), because charging all up front meant they were not being transparent about their ACTUAL costs and profiteering because the bulk of users simply switch rather than need to pay for a ‘cease’.

Some ISPs push up the charge from Openreach (eg to 50 pounds) while others simply absorb or pass on the Openreach charge (and customer pays the VAT, of course).

(*) not so many people are currently using FTTP which you mention, and I’ve not investigated the connections involved with respect to speech. There’s an obvious difference in getting the fibre to the property, and some of the costs involved are to do with covering the different equipment (as you said at 09:15) and the research and development costs.

I’m still looking to see what retail prices are available for FTTPoD which starts at £1500 (and could be much more), for setup and minimum of 36 Months contract. Openreach ‘wholesale’ charge to the ISPs at £99/month … that’s for the 330 Mbps service, but ‘retail’ prices with VAT on top are likely to be considerably more, so it will be very fast, but also rather expensive 🙂

John Ward says:

Peter – Thanks ever so much for all that detailed background information.

NFH says:

ISPs should publish three sets of figures (downstream/upstream for each):

a) Maximum theoretical speed of the broadband technology used (e.g. 24Mbps for ADSL2)
b) Likely connection speed (which for ADSL will be lower than (a) and dependent on the prospective customer’s phone number)
c) Likely bandwidth throughput at peak times after deducting 10% overheads and any reduction in speed as a result of contention ratios

When advertising generic headline prices, ADSL ISPs can only advertise (a); they can’t advertise (b) or (c) until the prospective customer submits their phone number.

rarrar says:

Exactly NFH

For Fibre (FTTC) most ISPs offer 2 packages with the max speed limited to 70Mbps or 38Mbps, although the connection in both case is capable of faster if you are close to the cabinet.

For those on fast connections c) bandwidth congestion is probably the most important factor and is affected by how much you pay !

wavechange says:

I see the Which? campaign is covered by the BBC at present:

“The Internet Service Providers’ Association said that it supported efforts to “boost transparency” but added that it was not always easy to work out what speeds consumers would receive.

“A variety of factors can affect speeds outside of an ISP’s control, including customers’ internal wiring and equipment,” said Ispa secretary general Nicholas Lansman.

“As speeds can involve factors beyond the control of the ISP, we would caution against too restrictive rules.”

Of course speed is affected by internal wiring and equipment, but even with a new router and computer connected directly to the master socket, many will not achieve anything like the speed suggested by the ISP.

dieseltaylor says:

Honestly there is a lot of guff here, and fortunately also some sense here.

Firstly no one has mentioned downloading from very popular sites or ones which have limited capacity where it is irrelevant what speed your ISP provides.

Secondly contention rates are always going to apply unless soemebody is putting in massive optical pipes.

Thirdly the ISP’s are quite correct that what customers do will have an effect on speeds and usability. Firstly wireless carries an overhead that can slow the data flow. The information you download may be in small blocks/file or you may be streaming data.

I have a neigbour who parked her wirless router in a cupboard under the TV and had the doors closed – and complained she could not get a good wireless signal upstairs.

SO please lets look at the whole picture of why you can get poor download rates to enable people to appreciate that in fact there are reasons other than the ISP that may reduce download speeds.

This from a dedicated site April 2013

” Further, according to Ofcom’s code of practice:
27. For services such as cable broadband where the main cause of disparity between headline speed and actual throughput speed may be network capacity limitations, ISPs must provide all consumers as early as practicable within the sales process, and in any event before consumers are asked for personal financial details, with information that actual throughput speeds during peak hours may be lower than at other times of the day where this is material. Specifically, where peak time speeds are likely to be more than 10% below the headline speed of the service, the ISP must indicate the throughput speed that is likely to be achieved during peak times.”

I suspect that this is not being carried out in the sales process or customers forget it has been mentioned. In any even it would seem very sensible that where a contract is arranged that the likely speed, and the reasons that it may not be achieved due to site constraints, or the user borking his equipment , is provided to the customer. This will be the basis of the contract and should include a choice of download speed testing sites to which te customer can check the suggested speed is being reached.

BTW I suspect in the advent of several heavy users in the neighbourhood and a lot of wirless use even that figure may be affected. Not to mention:

“There is an additional problem in newly built homes where builders have used plasterboard to build internal walls. They often use aluminium foil backed plasterboard to improve insulation which acts as barrier to the Wi-Fi signal, forcing the signal to bounce around rooms through doors or ricochet up and down the stairs, which can lead to significant loss of signal strength. “

wavechange says:

Good points. If you complain to an ISP you are likely to be told about these issues, and many of are already aware of them. It’s a good idea to restart the router and make a direct connection to a modern computer and making speed measurements over a number of days before registering a complaint. Hopefully most users are aware that popular sites can provide slow access at times.

The ISP is aware of relevant issues such as contention ratio and throttling, but what the customer needs to know is what to expect. As I mentioned earlier, my ISP provides a minimum and maximum speed that I can expect, and it is rare that it is lower.

PeterM says:

Good point about the plasterboard. I was helping a charitable group on an army base a while back and there were excellent signals from all the (soldiers’) homes OUTSIDE the building, but the wireless router in the room NEXT DOOR could not be “seen” (not just on a laptop, but the wireless printer, mobile phones, and so on). Had to install a cable so the new ‘wireless’ printer could be reached, because wireless was simply never going to work in that location.

The building was divided into a number of medium size rooms and I can only assume the plasterboard backing (for energy efficiency, it’s possibly a regulation for building works now) was the cause, not anything ‘special’ about the rooms / building.

wavechange says:

The majority of complaints to ASA about advertising of broadband services have been upheld: http://www.asa.org.uk/Rulings/Adjudications.aspx?SearchTerms=broadband#results

PeterM says:

In connection with broadband speeds and how accurate a prediction the ISP can be, I think the ASA have sometimes shown themselves to be missing the technical issues. I nearly called them fools but it’s a lack of genuine understanding, most likely.

PeterM says:

** Zoe ** – can I draw your attention to just-released figures from ThinkBroadband on their blog?

There’s a summary table on ISPreview at http://www.ispreview.co.uk/index.php/2014/11/broadband-download-speeds-top-20-fastest-uk-isps.html

I think the figures show just how complicated the statistics are, also gives a hint at the range of speeds, so there is no really simple way to class ISP “speeds” and the “up to” is an honest way to indicate what the ISP is aiming to offer (ADSL ‘up to 8’, ADSL 2+ goes higher, etc) and the individual speed is VERY MUCH down to postcode and phone number.

While it may be the case that for most streets, every phone (on BT) is connected to the same cabinet, the stats show that anything up to 3 or 4 cabinets in nearby streets can be used for a single postcode. The different cable lengths are one factor, and whether a cabinet is enable for FTTC is another, so in one street, some might get only 2 Mbps while others may be able to get “up to 24” and others “up to 76 Mbps” simply because of the wiring done maybe 40 years ago, to different cabinets in the area.

The ThinkBroadband blog posting, using data from hundreds of actual speed tests, can be found at http://blog.thinkbroadband.com/2014/11/october-2014-speed-tests-a-full-round-up/

Zoe Lederman says:

20 November 2014

Thank you for sending over those statistics, PeterM. You’re absolutely right. The statistics are complicated as speeds vary considerably. Regardless of whether this is due to a home’s distance from the telephone exchange, network capacity or too much traffic, I think the principle is the same. Why should people be enticed by speeds that they’re very unlikely to get? That’s why we think ads should be more transparent so that people can easily compare deals.

maxwild says:

15 December 2014

Obviously comparing Broadband offerings is so much more complicated that, say, TVs or dishwashers.
I have always used Thinkbroadband for speed tests but having read this and other conversations about speed I tried the speed test on my ISP, TalkTalk.
Thinkbroadband has consistently reported 6,5 or thereabouts for months as it did this morning. TT’s check showed 7.8 and suggested that this is slow as I should get between 8.2 and 17.
It suggested various actions including powering down the router – still 7.8!
So I rang TT, as they suggested, and after a few further automated checks I was put through to a human who ran a test and told me that I was getting over 9!
Of course I asked how this might be and he directed me to Speedtest.net where to my amazement 9.2 was returned. Leaving aside that TalkTalk obviously use two different measures this means that 3 tests at approximately the same time report 6.5, 7.8 and 9.2.
Does Which? do physical checks or does it rely on surveys? If so does it take account of different testing methods, postcode or distance to exchange and what sort of connection ADSL, ADSL+, medium or fast fibre.
I tend to agree with PeterM that if I am told between 8.2 and 17 then I have little to complain about if I get 9.2.
What I really want to know is if I moved to Utility Warehouse with 4 stars for speed compared to 2 with TT would I see a noticeable difference? If not why has UW got twice as many speed stars as TT?

dieseltaylor says:

Great to have some serious detailed input. Unfortunately rather than being included in an article it requires a passing reader to trawl through all the comments : (

malcolm r says:

dieseltaylor, one of my gripes about Which?’s approach to some topics is they try to whip up emotions on something they purport to be unjust, but do not give all the facts. So many are not able to make the rational judgement necessary (perhaps a relevant conversation is the Dr Grimes one – simply because it makes a plea for evidence to support claims allied to critical thinking). This approach also affects the validity of some Which? campaigns.

I’d like to see no more populist bandwagons promoted in the way politicians do, but reasoned argument and complete facts so we can make a thoughtful judgement.

My campaign would be to remove vat from energy; it is an essential purchase like food and rent, which do not attract vat. I expect that would attract huge support. However, I would also have to include the question of where would the lost tax be replaced from. It has to come from us. Where would that realistically be?

PeterM says:

to be fair, the figures were only shown today, so awkward for Which? to have them beforehand 🙂

wavechange says: