On first inspection, Ofcom paints a rosy picture of the state of mobile broadband in Britain. O2 (somewhat surprisingly, given our past real-world tests) tops the charts with average speeds close to 3Mbits/sec, with only Orange customers looking like they should find a new network.

However, examine Ofcom’s testing methodology more closely, and it becomes clear that those chart-topping 3Mbits/sec speeds are likely to be far higher than the average customer will receive.

3G only

Ofcom tested two different types of speed test: an automated “network capability” speedtest conducted by well-known testing firm Epitiro, and a more real-world test with a panel of 1,179 consumers.

The Epitiro test was conducted using static probes from numerous sites dotted around the country (locations pictured below).

However, as Ofcom admits, those sites were cherry picked, with only those offering the best 3G/HSPA speeds across all five networks making the cut. “More than 160 of the 600 candidate sites were surveyed to determine what coverage of 3G/HSPA services were provided by each of the five MNOs [networks],” Ofcom’s report states. “Sites where good 3G/HSPA coverage was available for the majority of MNOs were prioritised for deployment of probes.”

Not surprisingly, therefore, the overall average speed recorded by the static probes – 2.1Mbits/sec – is considerably higher than the average speed recorded by the consumer panel – 1.5 Mbits/sec – which includes both 2G and 3G speeds.

Yet, for reasons I cannot fathom, Ofcom has chosen only to publish network-specific average speeds using the artificial Epitiro tests, and not the real-world consumer tests.

Download tests

Ofcom/Epitiro didn’t only cherry pick the download sites – the test they used to measure download speeds also gives a highly unrealistic measure of performance.

Before I go any further, let’s all put our cards on the table: any type of benchmark, be it the PC Pro Real World Benchmarks or Ofcom’s mobile broadband tests, all involve a degree of artificiality. Benchmarks are designed to test the relative performance of one product/company against another using pre-determined metrics – they’re not necessarily indicative of actual consumer experience.

That said, the methodology used by Ofcom/Epitiro to measure web page download time is so far removed from the real-world experience as to be almost worthless, in my opinion.

“The webpage download test involved measuring the time taken to download the HTML ‘skeleton’ of three popular UK websites,” Ofcom’s report states. “The time taken to download the associated media assets, such as images and graphics were not included.”

Consequently, Ofcom admits that “the web page download times measured may be significantly faster than the time it would take to download a full web page with all images”. You don’t say.

Ofcom claims it would take an O2 customer an average of 1.5 seconds to download a web page – I’ve never seen page-loading times as fast as that on any mobile broadband network. In fact, you’re lucky to see those kinds of speeds on ADSL.

Image compression

Ofcom claims one of the reasons it decided to omit images from its tests is because some networks compress website images, while others download full-resolution photos, potentially skewing the results. A fair point.

The two networks that compress images – O2 and T-Mobile – come first and third respectively in Ofcom’s webpage download tests.

Yet, in my experience, the networks that compress images usually offer the worst page-load times, as the process of compressing the image on their own proxy server often takes longer than downloading the full-res image. In fact, image compression is usually the first thing I switch off (if the network even lets you).

So, by failing to take into account the bulk of page content (images are obviously far more data intensive than text) and the potential impact of image compression, Ofcom’s page download times should really be taken with an enormous pinch of condiments.

Worthless tests?

Does that mean Ofcom’s entire mobile broadband tests are baseless? Absolutely not. They are highly detailed tests, with a lot of very useful information in there for people considering a mobile broadband contract.

However, if you think you’re likely to regularly get 3Mbits/sec and page download times of 1.5 seconds, I suggest you’re sorely mistaken.

For years, fixed and mobile broadband providers have used the term “unlimited” to advertise services that are anything but.

We’ve moaned about it for years, and last month even our normally docile telecoms regulator said the term “unlimited” was being abused.  “There are people offering unlimited packages that contain a fair-use policy that means what you are getting is not unlimited,” said Ofcom chief executive Ed Richards. “If you are claiming unlimited then it needs to be unlimited.”

It seems the industry wasn’t listening. New data tariffs are still being advertised as “unlimited” even when they have specific download caps.

The Advertising Standards Authority (ASA) has been conducting a review of broadband advertising, but frankly, we’re tired of waiting for this weak-kneed, self-regulating body to get its act together.

So, from now on, whenever we see a new tariff being advertised as “unlimited” when it patently isn’t, we’re going to add it to our blog of shame.

T-Mobile has the misfortune of being top of our list, simply because it’s the first network we’ve noticed to launch an offending tariff since the Ofcom chief declared war on the term “unlimited”, but we’re certain there are many more.

So we need your help. If you see a fixed or mobile operator advertising a new “unlimited” tariff that has strict limits, let us know on comments below, and we’ll name and shame them too.

THE “UNLIMITED” BLOG OF SHAME

TARIFF: T-Mobile tariff for iPad 2 (new customers)

WHAT IT SAYS: £25 per month including 1GB Anytime data, 1GB ‘Quiet Time’ data and unlimited Wi-Fi.

WHAT IT MEANS: Wi-Fi usage is subject to a 10GB per month fair usage policy.

DATE LAUNCHED: 25 March 2011.

(Update: A T-Mobile spokesperson told PC Pro: “Blocking providers from using the term ‘unlimited’ when a data cap is applied is something which Ofcom is currently considering, but hasn’t yet put in place. There are a few different options which it is still weighing up, and this is just one of them.”)

The news that major ISPs are on the verge of signing up to a Broadband Stakeholder Group code of conduct on net neutrality and traffic management might sound like good news for consumers, but what will it do for the net neutrality debate?

The issue of how ISPs treat packets of data from various sources, and whether they can prioritise some websites over others if they have been paid for express delivery, has been hotly contested and there is a possibility that these guidelines will settle a dispute that regulator Ofcom has resolutely distanced itself from.

Indeed, in the absence of any higher authority there is a danger that the BSG guidelines could be seen as de facto regulations on how ISPs can approach net neutrality and traffic shaping – largely because the BSG is, it claims, “the UK Government’s leading advisory group on broadband”. It’s even part funded by the Government.

The group was set up to act as a “neutral forum for organisations across the converging broadband value-chain to discuss and resolve key policy, regulatory and commercial issues”. So in many ways it’s well placed to proffer an industry standard, but there remains a nagging doubt about the BSG’s jurisdiction here.

The very people involved in building this code of practice are the ones that have the most to benefit from a relaxed approach to net neutrality

The BSG is, in effect, little more than an industry body, representing at least in part the needs of the powerful players within that industry. So this code could easily be seen as the industry making up its own rules to suit its own requirements. It’s like asking Paddy Power and Ladbrokes to police match-fixing in cricket or the tobacco industry to decide whether cigarettes are a health issue.

When we asked the BSG who had worked on this project it said it had “discussed it in detail with the major ISPs and mobile carriers”, which makes it sound like the BSG was the driving force behind the talks and that the industry was brought into line by the group.

Dig a little deeper, though, and you could question whether the guidelines weren’t actually drawn up by the ISPs themselves and merely rubber-stamped by the industry body.

Surely not? The BSG is an independent forum, right? Well yes it is, but someone still has to pay the bills. A look at the sponsors of the group shows that although it’s part funded by the Department for Culture Media and Sport, among its other key financiers are 3, BskyB, BT, Orange, O2, TalkTalk, Virgin Media and Vodafone.

The very people involved in building this code of practice are the ones that have the most to benefit from a relaxed approach to net neutrality.

As we reported last year, there is already concern that Big Business is being given far too big a say in the net neutrality debate, that governments should be taking a stronger stand for the rights of consumers.

When a group that is funded by the industry and Government starts issuing edicts about how ISPs should treat traffic management and net neutrality, it’s easy to suspect that those Big Business concerns are about to trample consumer opinion underfoot.

“We don’t think it’s fair that the many should subsidise the behaviour of the few, and we think that we have a responsibility to our customers to address this kind of imbalance,” Dunne stated.

I’ll have you know that I’m the proud holder of A-level maths (grade C), and so I’ve dug my calculator out. By my reckoning:

O2 had 26 million customer accounts at the start of 2010, so it has 26,000 heavy data users

26 million x 200MB = 5,200,000,000 MB total data usage across the network per month

5,200,000,000MB  ÷ 3 = 1,733,333,333MB per month used by the 26,000 heavy data users

That means the average heavy data user consumes a staggering 66,666MB (so around 65GB) per month

Quite how anyone manages to download 65GB per month over a 3G connection is beyond me. You’d have to be running the connection round-the-clock. And indeed, when I put that scenario to O2’s press office, the spokesperson said that’s exactly what’s happening.

But instead of punishing the few to protect the many, O2 has done the exact opposite: it’s put a 500MB cap on previously “unlimited” accounts to ward off the data hogs.

If the rapacious appetites of the minority was causing O2 such a problem, why didn’t it impose a ceiling of, say, 5GB a month? By my calculations (A-level maths, remember) that would still wipe almost a third off O2’s total data traffic and only potentially infuriate a relative handful of its customers.

Instead, 26 million customers are now going to have to keep a careful eye on their data consumption, or risk being hit for excess charges of £5 per 500MB.

O2 insists that 97% of its smartphone customers will be unaffected by the 500MB cap and that the new charges will allow it to better plan investment in its network. “We have invested £½ billion in our network over the last two years; £10bn to date,” an O2 spokesperson told me. “In November 2009 we unveiled plans to accelerate our network growth, totalling £100 million over the coming year. This includes building 1,550 new sites across the UK by the end of this year.”

That’s all well and good, but let’s not pretend  an enormous data bill is destroying O2’s finances. O2 owner Telefonica recently announced annual profits of €1.6 billion, with its earnings statement paying tribute to the enormous growth in revenue from mobile broadband services.

O2’s CEO describes these changes as “fair and transparent”. I’m willing to credit him with the “transparent” bit.

One of the aforementioned cons of mobile broadband is image compression – a process where the networks water down the images on websites to conserve bandwidth. The end result is that sites such as the BBC homepage look as if they’ve been dipped in the bath, and in my experience, the compression barely saves any time at all on page downloads.

Many networks allow you to switch the compression off if you wish. Virgin Media doesn’t, on the rather dubious premise that it’s helping customers stay within their data download limits.

However, there is a sneaky way to beat the Virgin image washout, which I accidentally stumbled across during my tests. Virgin piggybacks on the T-Mobile network, and if you download T-Mobile’s Web’n'Walk Accelerator software, you’ll find that it can be used to adjust the compression on Virgin, too.

Instead of courage, creativity and innovation, the mood coming from Britain’s broadband leaders was complacency, resignation and a weary confession that we’re “still going round the same issues time and time again”.

The half of the country that’s connected to fibre provided by Virgin Media, BT or any number of local projects can almost certainly look forward to download speeds of 40Mbits/sec plus in the next few years. But what about the other half – the half living outside of the big cities that are already struggling on sub-standard connections?

Perhaps impetus for a nationwide next-gen network would be greater if the industry was willing to admit that our current infrastructure just isn’t good enough. But the complacency of BT’s director of public affairs, Tim O’Sullivan, was staggering to behold. Confronted with statistics that show Britain’s current position in the international broadband league is “more Championship than Premiership”, and a second report from Cisco that claims Britain’s crumbling broadband network could threaten our future competitiveness, O’Sullivan responded with platitudes bordering on arrogance. More than 99% of the country already has access to broadband, he retorted, 40% of which have access to “up to” 20Mbits/sec connections on BT’s network. “People should take stock and have a dose of reality in these debates,” he snapped.

If he wants a dose of reality he should talk to the Country Land and Business Association (CLA) and its tens of thousands of members who can’t get a broadband connection worthy of the name. He should talk to the small business owners who are forced to relocate because they can’t get a stable ADSL connection. He should talk to the people who are seriously considering moving to mobile broadband – and its average connection speed of 0.87Mbits/sec – because it’s still faster than their landline.

Ofcom’s lapdogs

Yet is it any surprise that BT rests on its laurels, when even our supposed consumer champions are so desperately meek? “2Mbits/sec is good enough for today, but it’s absolutely not enough for tomorrow,” Anna Bradley, chair of the Ofcom-sponsored Communications Consumer Panel told the audience. Sorry Anna, but 2Mbits/sec isn’t good enough. It isn’t good enough to watch the HD streams from BBC iPlayer – a service the vast majority are already paying for through the licence fee. It isn’t good enough for many people to work from home, because of the deplorable upload speeds. It isn’t good enough in modern multi-PC households, with two, three or more devices trying to connect to the internet simultaneously.  No wonder that a spokesman for Lord Corbett referred to the Communications Consumer Panel as the “Industry Backside Protection Unit”.

Hope for the future?

So what hope is there for nationwide superfast broadband? BT says there’s no business case to lay fibre in any more than 40% of the country; some others estimate that 60% may get fibre by market forces alone. But even the most optimistic predictions don’t go above that figure.

Mobile won’t step into the breach. “There isn’t spare capacity in mobile networks to augment fixed-line networks,” said Forrester analyst Ian Fogg. “Mobile next-generation access is about the twenty-four seven lifestyle – it won’t be an alternative to fibre or Virgin Media’s network.”

Which leaves half the country pinning its hopes on Government intervention – although few think Lord Carter’s broadband tax will raise anywhere near enough to fibre the nation, and the Tories have pledged to scrap it if they get into power next year anyway.

In fact, the best hope of widespread fibre might come from innovative start-up projects, such as Fibrecity’s cost-cutting solution of running fibre though the sewer network. The company’s Adrian Cook said Fibrecity would shortly announce another ten cities that would benefit from the technology, and laughed off suggestions that fibre deployment need cost anything close to the figures suggested by the likes of BT and the BSG. “[We can install fibre for] significantly less than £400 per home,” he said. “If I were anywhere near that, my chief financial officer would have something to say about it.”

Perhaps someone can put a word in, and get this man a job at BT.

For the last few days, I’ve had to “rely” on mobile broadband as I wait for my broadband connection to go live in my new house. The trouble is, it doesn’t work at all well. The first problem is reception: I don’t live in central London but in deepest Bucks, and that means I can only get a GPRS connection.

Or at least, I could. On Saturday last, I had a solid connection for an hour, and though browsing was a slow, awkward affair (anyone remember dial-up?) I managed to get a few things done, and make myself numerous cups of coffee in the spare minutes as I waited for pages to load.

But then, firing up my laptop on Sunday, I was faced with a series of error messages. And, as with all error messages, the recommended steps were of no use whatsoever. I’ve since worked out that it was due to two processes working at the same time (why doesn’t the error message say this is the likely cause?) but as the processes aren’t terribly easy to kill I only had success after a couple of reboots.

And today comes the final nail in my mobile broadband coffin. I’m sitting on the train writing this, with a Vodafone dongle sitting proudly in my machine, but travelling from Aylesbury to London I’ve barely managed a solid connection of longer than five minutes. GPRS network detected, it says. By the time it’s gone through its handshake procedure, we’ve shifted onto another network. GPRS network detected. Signal very poor.

The fact is, I can’t do anything like this. I’ve sent two emails and read five in the same time I’d have sent a dozen and read 50 normally. It’s a long way from being productive. And it’s unfair to single out Vodafone – I’ve had similar experiences with 3, and its software is even worse than Vodafone’s.

Perhaps one day all this will be solved, but bearing in mind how long the 3G network has had to mature it’s incredibly disappointing that the service is so frustrating and unreliable, especially for people on the move. If it hasn’t been sorted by now, will it ever be?

Here, for example, are two mobile broadband dongles from O2 (left) and BT (click here to read about BT’s new mobile broadband service). One causes me endless hassle on the train journey into work of a morning, while the other is painless. The difference? That little white piece of string that keeps the BT dongle’s cap connected to its body. 

Something as innocuous as a cap retainer might sound utterly trivial, but I’ve lost count of the amount of times I’ve ended up on my hands and knees, picking my way through the half-eaten bag of Doritos and discarded newspapers under the train table, trying to find the AWOL O2 cap. It doesn’t even fit on the other end of the stick!

The BT cap, meanwhile, remains firmly anchored to the stick, no matter how many hard-disk threatening bumps the train encounters.  How much does that little piece of string add to the cost of the device? A tiny fraction of sod all. How much difference does it my mobile broadband “experience”? A pretty sizeable one.

“I’m not a violent man but right now I would cheerfully hurt someone from BT,” Gorman writes on his blog. “In fact my sense of proportion has diminished to the point where I can’t work out if it would be in particularly bad taste to suggest that running Kris Marshall over again would be, well, satisfying. Probably.
Still, if he will advertise BT’s services…”

However, what piqued my interest was his superb suggestion of ISPs providing mobile broadband dongles to people whose landlines have given up the ghost: a courtesy connection in the same way garages provide a courtesy car when your runaround is being repaired.

Sadly, it seems BT (which, of course, flogged off its O2 mobile arm years ago) isn’t buying the idea, judging by Goram’s transcript of his conversation with the company’s “support” staff:

Do you want to help me solve the problem?
Yes.
You know those USB modems that provide mobile internet access?
Yes.
One of those would solve the problem wouldn’t it?
um…
Wouldn’t it?
… well…
It would wouldn’t it?
Yes.
So… can you, BT, provide me with one?
No.
Why not?
Because that’s not my department.
So whose department is it?
…um…
Could you put me through to the person who is able to make that decision?
I don’t think anyone can?
Of course they can. It’s possible. The things exist. Someone somewhere makes decisions.
I don’t think that’s how we deal with the problem…
I can see that. But so far you haven’t dealt with the problem. Have you?
um…
Have you?
Well…
It’s still not fixed is it?
No.
You know those USB modems that provide mobile internet access?
Yes.
One of those would solve the problem wouldn’t it?
um…
Wouldn’t it?
… well…
It would wouldn’t it?
Yes.

Until yesterday, T-Mobile had provided us with the larger Huawei E220 USB modem you can see at the top of the photo here. And to be honest, it was pretty ropey. Tim Danton described last week the trouble he had installing the device and the download speeds we recorded – even when sat upon the 6th floor balcony here at Dennis Towers – were distinctly underwhelming. Speeds were typically hovering around 300-400Kb/sec, placing T-Mobile well behind rivals such as Vodafone and 3.

Then we spotted that the company had released a new Huawei USB stick (dubbed web’n'walk Stick III) that plugs straight into the PC – near identical hardware to that used by Vodafone and 3, in fact. We badgered T-Mobile’s press office to send us a test unit and the results have been nothing short of remarkable. Download speeds are now hovering around 2Mb/sec – around five times faster than what we were getting with the E220 modem. Even on my train journey from Sussex to London the connection is reasonably solid, whereas before it was decidedly choppy.

If you’re thinking of signing up for T-Mobile’s mobile broadband, make sure you opt for the Stick III rather than the E220. If you’re still not sure what mobile broadband network to hook up with, make sure you read issue 166 of PC Pro, on sale 19 June, where you can find out how all five of the UK networks compare.