Lies, damn lies and advertised broadband data speeds
Ofcom has just published the results of the UK’s most comprehensive broadband data speeds research (August 2010). It revealed a growing chasm between the advertised speed and the speed actually experienced by customers. Packages that were advertised as offering speeds of up to 8 Mb/s were coming in at the following measured results (see Ofcom Web site for the full data):
AOL – 3.6 to 4.7 Mb/s
BT – 3.8 to 4.5
O2/Be – 4.3 to 5
Orange– 3.3 to 4.2
Talk Talk – 3.6 to 4.3
Sky’s package advertised speeds of up to 10 Mb/s – a whole 2 Mb/s more than the above ISP’s up to 8 Mb/s packages. But Sky’s measured result, according to Ofcom, was only 3.9 to 4.9… most of the extra 2 Mb/s mysteriously vanishing somewhere along the superhighway…
The situation that Ofcom and the Industry find themselves in reminds me of the airline industry. A few years ago I was doing a lot of flying aroundEurope. Over this period almost every early morning business flight out ofHeathrowAirportwas consistently around 1 hour late. On one flight I happened to find myself sitting next to an official from the UK Civil Aviation Authority. I remarked to him about this consistently poor track record. Partly in jest I suggested that it might be a lot better for everyone if the airlines reprinted their morning time-tables with 1 hour added to every departure time…then all the airlines would leave on time!
The reply from the official was quite revealing and it has a bearing on the rather exaggerated broadband speed claims from most ISP’s. He said that all the airlines did their market research and this told them that business customers wanted to arrive at their destinations for 10.30 meetings which made it absolutely clear what time the flight had to leave Heathrow to be attractive. This left all the competing airlines demanding exactly the same take-off slots. Short of 5 aircraft all going down the same runway at the same time – that was impossible. But the marketing departments knew it was commercial suicide not to offer the ideal departure time in their time-tables…and there was a chance (a rather small one) that a particular flight might acutally leave on time. So the game for the rest was to blame air traffic delays…
This is where most of the broadband ISP’s find themselves today but compounded by the fact that one ISP, Virgin Media, is not in the same boat (or using the same airport to continue the analogy) as they enjoy the legacy of a superior local drop cable – a coaxial cable rather than copper wires. Their service advertised as “up to 10 Mb/s” actually delivers 8.6 to 9.0 Mb/s. So for the remaining competitors to only advertise around half these speeds would be commercial suicide in cabled areas…even if that is the reality for most of their customers for most of the time?
Ofcom has discussed this issue with the Advertising Standards Authority (“ASA”) and the Committee on Advertising Practice (“CAP”) are currently undertaking a review in this area. But the speeds achieved down an IP network are a statistical outcome of a large number of complex traffic events… …and it is statistically possible, in a momentary lull in IP traffic, that a user might actually get the advertised speed…but for how long and how often? In addition it is simple good economics on an IP network to have a certain degree of overselling of capacity (known as the contention ratio). It all comes down to probability and statistics and we all know about lies, damn lies and statistics…a long standing tool of the trade of the advertising industry. So I would not be holding my breath that the Committee on Advertising Practice will be second guessing the industry on such a complex technical issue.
Ofcom suggest the industry itself should agree a “standard currency” for describing broadband data speed performance, similar to those in other industries (for example, APR in financial services, and MPG in motoring). But somehow I find it hard to imagine BSkyB and Virgin Media sitting together and agreeing a measure in the very near future that reveals rather too clearly that one ISP’s “up to 10 Mb/s” product appears technically almost twice as good as the other in terms of customer experience.
Where I suspect we are likely to finish-up is Ofcom continuing its research, publishing the results and continuing to admonish some ISP’s each year for not delivering the speeds that their advertising is leading consumers to expect. With a bit of finessing this may actually be the most useful outcome in both informing consumers of “the best data speed buy” and in this way cajoling the individual companies to invest more to improve their Ofcom annual measured rating. The DSL copper wire based delivery systems are capable of matching the Virgin cable performance…it just requires a substantial investment to shorten the copper wire runs (fibre to the street cabinet) and unblock the current middle mile congestion. It might also be useful for Ofcom to break their future results into what is happening inside Virgin Cabled Areas (around 40% of the population) and outside since I suspect investment in up-grading BT’s copper wires is likely to be targetted where there is the most competitive presssure…
If this is the likely outcome on this issue then we all have to entreaty Ofcom, in an era of budget cuts, to keep this bit of outstandingly valuable research fully funded…