U.K. Broadband: Lies, Damned Lies and Statistics

Samuel Clemens once quipped that 'there are three kinds of lies: lies, damned lies, and statistics." So the U.K. Office of Communications says real-world end-user average peak throughput is about half the "advertised" broadband speed users pay for. 

On the other hand, Ofcom also notes that average speeds run between 81 percent to 85 percent of the "advertised" speed. 

In fact, subscribers on services promising 2 Mbps or less get those speeds about 91 percent of the time. Since most broadband subscribers in the U.K. are on lower-speed tiers of service, the national average speed delivered is about 85 percent of the maximum line speed.

So whether performance is "good" or "bad" is a statistical matter, depending on whether peak throughput at the peak congestion hours are examined, compared to average performance across each day, week or month. "Peak" performance also hinges on matters beyond a service provider's direct control, such as the state of in-home wiring and capabilities of in-home end user equipment. 

That said, the difference between "peak" throughput and "average" is directly affected by the fact that access is a shared resource, in the access network, in the aggregation network, on the backhaul networks and at the servers users are trying to communicate with. 

The actual throughput received by its national panel of testers was 3.6 Mbps in the 30 days beginning October 23, 2008. That throughput represents 49 percent of the average ‘headline’ speed (7.2 Mbps) and 83 percent of the average maximum line speed (4.3 Mbps), Ofcom reports. Consumers on the most popular broadband headline speed package, advertised as offering  "up to’ 8 Mbps," received an average actual throughput speed of 3.6 Mbps, about 45 percent of the headline claim, and they had  an average maximum line speed of 4.5 Mbps, representing 56 percent of headline speed.

About 20 percent of testers on the 8 Mbps package received an average speed of less than 2 Mbps. Still, about 83 percent of respondents say they are "happy" with their service, while 21 percent report they are dissatisfied for some reason. About 16 percent express dissatisfaction with the "value for money" they receive and 13 percent are unhappy about service reliability.

About 28 percent of users were unaware what the advertised speed of their connection was. Rural consumers on "up to" 8 Mbps packages received average speeds 13 percent lower than their urban counterparts.

To be sure, Ofcom notes there are many reasons why throughput might be slowed. Congestion on the wider internet, loop length, the condition of the access cables, poor home wiring, absence of filters, computer clock speed or router specs can degrade performance.

Honesty in advertising is an issue, of course. The issue is that broadband access is a shared, best effort resource. There will be times when any single user actually will experience the full advertised throughput. Most of the time, likely not.  So long as users use "maximum speed" and "cost" as the key criteria to compare providers, it is not likely the advertising verbiage will change. 

The other issue is how to describe "average" speeds, across different provider networks, in ways that are consistent and meaningful, since the actual end user experience always will differ for all sorts of reasons, some not under the direct control of the access provider. 

$10 iPhone Tethering?

The rumor that AT&T is considering both Apple iPhone tethering (allowing it to act as a PC modem) and an incremental $10 charge over a standard data plan, would be a smart move, in the right direction. 

Since AT&T expects virtually all consumer devices to be capable of wireless broadband connection, some of us already have been figuring out what that means, cost-wise, for the devices we support, keeping in mind that some of us support a whole family's requirements.

It is self-evident that this future will have a tough time becoming material reality if present pricing for mobile broadband remains where it is. There is no way most parents would be willing to pay $30 to $60 per connected device to participate in such a world, where a dozen to scores of devices might plausibly need to be connected. 

The only way this idea really becomes a mass option is to create unified data access plans built on the notion of family plans, where all devices and people can share one bucket of access. Even fixed broadband penetration would not be where it now is if users had to pay a separate fee for each device, and each user, accessing a single connection. 

Windows 7: Faster, Longer, Fewer

"Windows 7 should boot more quickly, have longer battery life and fewer alerts,"says Steve Ballmer, Microsoft Corp. CEO. That would be nice. All three are present annoyances.

AT&T Mobility Bundles PC

Dell  and AT&T Mobility have launched a limited-time offer (ending Jan. 31, 2009) bundling 3G service with a PC, requiring a two-year contract to AT&T "LaptopConnect" costing $60 a month. 

The offer extends the common mobile phone offers that bundle discounted handsets with service to mobile PC service. The Dell Inspiron Mini 9 will cost $99 after a $350 mail-in rebate. 

Orders can be placed on Dell.com.

There are several obvious implications, some pertaining to service provider strategy and revenues, some pertaining to public policy issues. The service provider angle is that, as handset subsidies have boosted mobile subscriptions, so PC subsidies will boost use of mobile PC data plans. 

The public policy angle is that, to the extent there are users who want to use the Internet, but do not own PCs, this sort of bundling addresses their needs. To the extent there remains a gap between desire to use the Internet, and the means to do so, programs of this sort will address the problem. Some people do not want to use PCs or the Internet, and virtually nothing is going to entice them to do so. On the other hand, bundling access devices with service is a proven way to stimulate demand. 

Another angle is that this bundle moves us further towards a world when broadband access will be a "personal" service. Where in the past voice service was to "places," it now is to "persons." Where broadband access largely is to "places," this sort of plan moves us in the direction of broadband access to "people."

Frogs at the Bottom of Wells

Inevitably, all of us are paying quite close attention to all things economic these days. Just as inevitably, we journalists and bloggers cannot resist writing about it, and what it means for all manner of things, ranging from penetration rates of various devices, services and applications to levels of industry spending.

Crowd sourcing, as valuable as it is, also can be dangerous, though. The reason is the well known tendency people, and therefore markets, have to overshoot on both the upward and downward sides of any trend.

So one easily can take a poll of industry participants (largely on the sell side) and find dire opinions about the state of service provider spending (and therefore buying). Keep in mind an analogy: the frog sitting at the bottom of a well, and asked to describe "the sky."

By definition, all of us have limited visibility. None of us can see the whole sky.

And our view is obscured in several obvious ways. The overwhelming amount of spending in any country or market is dictated by just a few buyers. Some types of products are needed more, some are needed less. Big capital projects are needed, but end at some point.

Sure, most providers have some base level of maintenance-related capital spending that doesn't change much from year to year. But there tend to be waves of investment in the global communications business that ebb and flow.

There are times, such as 1998 to 2000, when spending, in absolute volume, climbs, and periods such as 2001 to 2003, when the percentage falls. Capital spending fluctuates, for all sorts of logical reasons.

Then there are the other obvious visibility-limiting issues, based on which customer segments and which product lines one sells. Some segments do better, some worse. Some products are necessities, others can be postponed.

All of us will have a tendency to attribute virtually any shifts to the downside as caused by the economy. That isn't always true. Nor is it true that service provider revenue actually is falling. Through 14 months of recession, service provider revenue has grown virtually across the board, for every segment, though there are market share shifts and secular changes in demand.

Executives are being prudent, to be sure. Investors and investor advocates demand that. But we all have the visibility of frogs, who think the sky is a relatively small blue circle. So we will overshoot, as we always do.

Without dismissing in any way the obvious issues the industry confronts, do not mistake your own view for the whole picture. And do not make the mistake of believing that any present trend can be extrapolated into the future on a linear basis. By definition, there are turning points. The year 2000 was a turning point. So was 2003. It appears 2008 will mark a turning point. Another is coming.

If your business will last more than several years, you have to spend some time looking for the next turning point. It is hard to do. But it is coming.

Global communications infrastructure spending has a "float" level, generally a percentage of revenue, to which the industry is trending after a buildup leading to 2000, a brief lull, then a wave of infrastructure spending largely driven by broadband. At the moment spending is drifting back to "maintenance" levels.

Another wave is coming though, based largely on the fact that broadband multimedia networks must spend much more discretionary capital on consumer premises equipment. As more services are turned up, there is more spending on CPE. That's a secular change strictly driven by revenue opportunities.

Also, Internet-delivered video, at some point, is going to drive more revenue for service providers, as opposed simply to application providers. As that happens, more investment in access networks will have to be made. The precise timing will depend on end user uptake and therefore new revenues.

But tier one service provider performance in 2008 suggests the process has moved significantly further than even many tier one executives had expected in 2006, for example. To cite but one example, where many had expected new data revenues only to offset voice revenue losses, many carriers now find data revenue growth is outstripping mere replacement of lost voice revenues.

Look for the next turning point, even as you manage for the present circumstances.

Boucher Replaces Markey: Expect Changes

Congressman Edward Markey, a key proponent of net neutrality, will leave his position on the committee that deals with telecommunications regulation to chair the House Energy and Commerce Subcommittee on Energy and the Environment. He will be replaced by Congressman Rick Boucher, who takes on the chairmanship of the Communications, Technology and the Internet Subcommittee of the House Energy and Commerce Committee.

It is worth noting when key communications regulators change seats, since regulators are a primary force in the creation of permissible business models and the potential profitability of communications business models. If his past actions are any indication, "net neutrality" is going to get a lot less attention, rural broadband much more.

Boucher is likely to support plans to tie universal service support to broadband, not voice. That could have positive investment implications for rural telcos and even for some tier one providers. Qwest, for example, has large rural service areas where it might benefit from increased support for rural broadband.

"The indication right now is that the Obama administration will be thoughtful," says Qwest CEO Ed Mueller. So support programs for rural broadband could change. Qwest favors a bidding process for any new government support for building rural broadband facilities, a process it believes it can win. "But we think we'd get a decent return on that," Mueller says.

The other structural change that would help Qwest is if USF funds were awarded on a community-by-community basis, not on a statewide basis. The reason there is that Qwest operates in many states where it serves both urban communities and lots of smaller rural communities. Obviously, that formula restricts Qwest from getting USF support to serve a large number of rural communties .

"More broadband support would be good for Qwest," says Mueller. "We just want to bid on it."

Using "Password Manager"

I've been testing a password manager program, "Password Manager," created by Large Software (www.largesoftware.com), which earlier released PC Tune-Up. Now, some of you may not think a password manager is a useful thing, so you can skip to another post. Personally, I live on the Web for professional reasons, and there no longer is any easy way for me to remember all my user names and passwords.

Some of you may be diligent about storing all your user names and passwords someplace, but that gets to be a chore, not to mention a security risk. Those of us who travel need our identities with us on the road. Retrieving them from a secret location in our offices or homes is not a convenient option. Sticky notes are worse.

Password Manager encrypts each saved password and also protects the program by offering a master password, keeping all the information stored protected and secure. For obvious reasons I have not tested that particular feature. I also use fingerprint readers to protect access my machines, so the additional level of protection is comforting, I will say that.

Password Manager is built to automatically recognize when a user is returning to a site for a login. For new accounts, the system completes the user’s login information in order to eliminate cumbersome set-up forms. Keystrokes are concealed to avoid keyloggers or spyware from stealing sensitive information.

The application pretty much runs after you download it. I did experience a bit of wondering whether it was working at first, as I was using the Google Chrome browser. I figured out what was going on after I switched back to Mozilla Firefox. Password Manager also works on Internet Explorer, of course.

If you decide to change a password, the program gives you a prompt, asking if you want to proceed.

The login information and passwords are available only when the password storage database is unlocked by an authorized user. In my case, the fingerprint swipe seems to do the job, so I can't speak to whether one has to enter the master password to activate the vault.

Password Manager is said to protect users from keylogging (the unauthorized monitoring of your key strokes by a third party). Again, I believe this claim, though I haven't tried to verify it by hacking my own machines.

I recently spoke with a buddy who uses a different password manager and was unhappy with it. Except for the fact that I have to remember whether I am in Chrome or Firefox or Internet Explorer, I haven't had any problems at all. Add support for Chrome one of these days and I'll be even happier.

Each of you will have to decide whether it is worth $30 to automate your password entry chores. Speaking for the skin on my fingers, it is quite useful. I use fingerprint readers so it is important there be recognizable skin on those fingers!

By the way, I just ran a PC Tune-Up scan on a machine that hadn't been scanned before and the software found 266 problems, about half of the "high priority" sort and about half of the "medium priority" sort. I'm a dumb end user so I have no idea what all that stuff was, though they seem to be "invalid application paths."