if connected

How much worse does broadband have to get?
Here’s the dilemma I wrote about last year (and also here). ISPs need to get together and sort this or everyone loses out: lower per customer broadband revenues for ISPs, unhappy consumers, plus hostile articles in the press or on consumer affairs TV like Watchdog attacking individual ISPs, and slow broadband hindering Internet businesses offering Internet TV or other innovative services.
My concern is that too many in the ISP industry are using the quality of telephone lines as an excuse. It’s not.
Too often consumers’ telephone lines are operating at a much much higher speed than the actual broadband speed they are receiving.
Most consumers will not realise their telephone line is capable of more and is not the bottleneck, and just complain that the service is slow. ISPs pass the buck between retailer and wholesaler or blame technology limitations (that do exist but are frequently used as incorrect scapegoats).
Almost all consumers will find it impossible to decide which ISP is better/worse as all ISPs are opaque in their marketing of broadband packages.
Here’s an example of some speed tests(see the end). These results are using the official BT broadband speed tester on a DSL service retailed by a small ISP, but which based on a BT Wholesale business grade service.
The ‘IP profile’ of 7150 means the DSL modem is synchronised at 8128kbps or “8 Meg”, which is the fastest possible on a BT Wholesale-based service (for now). 7150kbps is the maximum actual speed the telephone line is capable of delivering (it’s lower due to error correction and other technology overheads).
But, the actual speed delivered is much much lower, as demonstrated by the speed test results.
The average speed over a period of a month of testing, across 10 separate tests, was just 29 percent of the 7150kbps theoretical maximum.
This is not an isolated example.
1/6 at 14:11pm
IP profile for your line is – 7150 kbps
DSL connection rate: 832 kbps(UP-STREAM) 8128 kbps(DOWN-STREAM)
Actual IP throughput achieved during the test was – 2084 kbps
1/6 at 14:22pm
IP profile for your line is – 7150 kbps
DSL connection rate: 832 kbps(UP-STREAM) 8128 kbps(DOWN-STREAM)
Actual IP throughput achieved during the test was – 1931 kbps
11/6 at 9.26am
IP profile for your line is – 7150 kbps
DSL connection rate: 832 kbps(UP-STREAM) 8128 kbps(DOWN-STREAM)
Actual IP throughput achieved during the test was – 2386 kbps

One year on from Carphone’s “free broadband” launch, the CEO apologises:

“I wish I knew a year ago what I know now, we could have saved ourselves and our customers much stress.”

and:

“Thank you again to everyone who took the plunge at the beginning, sorry if we caused you problems. If you haven’t joined the revolution yet – one year on we are a much bigger, wiser and more reliable supplier maybe it’s time you gave the Challengers a chance to prove that we will save you a great deal of money and provide you with a great product.”

Of course, after last year, the brand has suffered and Carphone will need to work hard to re-build it.
Read Jupiter’s analysis of the UK’s “free broadband” offers or read a case study on Carphone’s “Free Broadband Forever” launch and the impacts on customer service and retention.

On Tuesday, the net neutrality debate arrived at the UK’s seat of government. The debate centred on the industry view, not consumers (unlike Jupiter’s recent report), but was revealing nevertheless.
Despite the headlines of many articles I’ve read, Ofcom does not rule out action. It just believes, rightly, that new legislation is not needed to ensure a functioning Internet market:

Scott [Ofcom] also said it would not be wrong for an ISP to approach an application provider offering to guarantee service quality for a fee — unless that ISP had significant market power [my emphasis], in which case Ofcom could weigh in on the grounds of anti-competitive behaviour.

Net neutrality (or not) is a fundamental part of the business models of both Internet Service Providers (ISPs), and for any company offering products across the Internet.
Dougal Scott, Ofcom’s director of policy development on Tuesday as quoted by ZDnet:

“There is a very rapid increase of traffic on the internet,” said Scott on Tuesday, pointing to a “change in the nature of applications that people are using on the internet”, particularly time-sensitive applications like voice over IP. He went on to characterise the “all bits are equal” advocates as the “most extreme” fringe of the net neutrality lobby, and insisted that there were “real advantages to consumers in treating certain types of applications differently to others”.

This goes wide of the mark: It’s not giving different types of Internet usage different priorities, or quality of service, that is at the heart of the debate. It’s whether different companies’ traffic, or bits, are treated differently from other companies, or from an ISP’s equivalent products.
Will Skype be prioritised over Vonage by an ISP?
Will Apple TV be impeded, or degraded, by an ISP while Joost and Babelgum are not?
Will all of them suffer second class treatment compared with a BT’s IPTV or Orange’s Wireless & Talk VoIP services? (both of whom are ISPs)

The real issue, said Scott, was not the “traffic shaping” policies of ISPs, but the way in which those were communicated to the public. “If you use BitTorrent, how do you know which ISP to go to?” he asked, adding that there “needs to be greater clarity to the consumer on ISPs’ traffic-shaping policies”.

Yes, absolutely, it’s good to see some UK decision makers catching up with what I wrote about last year here If ISPs are to resist downward price pressure for broadband packages, consumers need to understand why they should pay a premium for one package over another, or one ISP over another. How will consumers know whether the broadband connection they are paying for does what they want?
For a consumer view, read Jupiter’s new ‘Net Neutrality in Europe’ report. This contains a segmentation based on attitudes to net neutrality and looking at the potential to monetize each group.

Interesting data on actual end user broadband speeds in this Australian government report.

On these figures, 62 percent of consumers connected at 512Kbps or slower, and a staggering 42 percent ‘enjoyed’ under 256Kbps, at the end of September 2006.

The impact, for online video services and other emerging broadband applications of low speeds, data volume limits, and other package constraints is immense and an area we’re tracking carefully in Europe. For comparison last time I tested it, Joost — formerly the Venice Project — really needs 512Kbps to 1Mbps broadband and an unlimited data volume to fly. More on this in several upcoming reports.

[update 6.02pm 27/7] This isn’t intended to be an attack on BT Wholesale. The issue here is about clarity and transparency in marketing broadband and is one *everyone* in the industry needs to tackle. More tomorrow.

Back in March BT Wholesale launched the line max services that offer speeds theoretically up to 8Mbps. Today, in the BT results BT Wholesale argue this has been successful and a stable product, I think the industry would beg to disagree.

BT Wholesale launched broadband max services in March 2006, providing the UK market with the highest stable speed broadband service across the widest national footprint in the world. Service providers are currently migrating broadband subscribers to BT Wholesale’s IPMax services, which deliver services at speeds of up to 8 Mbit/s. To date, more than 750,000 subscribers have been migrated to max services, aided by the availability of a new automated mass customer migration tool.

The vast majority of DSL broadband connections in the UK are based on BT Wholesale packages. If this wholesale product has issues it affects almost every ISP in the industry. Several ISPs are offering free re-grades back to the old flat rate broadband products, and one of the leading forums for savvy broadband users reports multiple problems with line max that are continuing.

BT Wholesale advised that the reasons a consumer might not be able to receive that speed was due to telephone line quality issues, ie a technical limitation of ADSL1 that is, correctly, outside of BT’s control.
In reality, speeds are being limited by other factors (as well) such as exchange congestion. Plus, BT have set their systems to include some apparently rather arbitrary speed throttling settings, which are intended to improve reliability but appear to create a lowest common denominator for download speeds and have caused a lot of teething problems.

The problems for the industry as I see them from this are:

1. Advertising standards issues with the 8Mbps claim. An actual 8Mbps download speed is impossible to achieve on ADSL1, so these products should not be advertised as ‘up to 8Mbps’. The fastest actual download speed is around 7.1Mbps.

2. Misrepresentation of the reasons why the maximum speed may not be achieved. An end users’ speed is not determined solely by the quality of their line at the moment. Yet, that is what most ISPs selling these services imply, see: Demon, Nildram, Pipex, Plusnet, Orange, etc.

Actual speeds appear related to congestion and internal BT network settings. While previous products were sold as “contended” with a stated 50:1 or 20:1 ratio of users to capacity. That has been dropped from the product description now. In practice, contention issues appear very common now, while they were rare before.
Additionally, those ISPs that use traffic shaping, or prioritization, on their own network may reduce usable speeds for consumers. For those ISPs, blaming the quality of a customer telephone line is a convenient decoy. This reduces transparency in the market.

3. The above will create further downward price pressure on the broadband market (see our forecast for broadband ARPU). If consumers think their line is the reason they cannot get higher speeds, which I suspect most consumers will think (as few will know to check their router sync speed and compare it with the actual download speed), then they are extremely unlikely to think it’s a good idea to switch to a more expensive ISP offering faster speeds. It makes it hard for an ISP focused at the quality end of the market to justify why a consumer should pay more for broadband; this puts further pressure on broadband prices which is not in the industry’s interests (as a whole).

4. Poor performance of BT Wholesale-based DSL will increase market concentration around the largest ISPs, reducing consumer choice between providers. If the BT Wholesale-based packages have too many problems it increases the need for ISPs to invest in LLU networks to be competitive. But that requires significant economies of scale to be worthwhile. Medium-sized and small ISPs will find it harder and harder to remain viable; unless the LLU players start offering a compelling wholesale service to compete with BT.

5. BT Wholesale will likely use the line max teething problems to justify their late ADSL2+ launch timescales. In today’s announcement, BT reiterated that the DSL2+ max products (up to 24Mbps) would be tied to the 21CN roll out and therefore start in January 2008. Be (Telefonica-owned) launched DSL2+ a year ago in the UK. The hidden jewel in Sky’s announcement last week was the mass market launch of its DSL2+ at up to 16Mbps for just 10UKP a month extra for existing Sky TV customers. This will increase the benefit to ISPs of using an LLU network, rather than BT’s. Elsewhere in Europe, ADSL2+ services similarly launched years ago.

This piece has a good summary of funding that the UK’s BT has received to increase broadband availability in rural areas. This is one of the (many) areas where Europe has quite different broadband dynamics to other parts of the world.

• PlusNet and Easynet announced that PlusNet will offer ADSL2+ services based on Easynet’s DSL LLU infrastructure (in addition to their existing BT-Wholesale based products). ADSL2+ will deliver speeds of up to 20Mbps (perhaps even 24Mbps), provided that the copper line is good enough quality.
• BT Wholesale announced the next step in their 8Mbps ‘line max’ services. A beta trial will go live this autumn but commercial services will not be launched until Spring next year. Compared with most of Europe, this will be quite late.
• The UK telecoms adjudicator reported today that he was disappointed with the progress made in LLU. This is hardly news to any of the LLU ISPs, but will put additional pressure on the new BT Openreach division which is charged with delivering ‘equivalent’ access to the copper loop for all providers.

Bottom line: Broadband speeds are about to surge in the UK, but those speeds are taking longer to arrive as a mass market product than many would like. ISPs involved in LLU are suffering additional customer service costs due to the imperfect LLU process. Consumers will have more choice of broadband product in the future but also more potential for confusion.