When the OECD first released highlights from its Broadband Report in early June, I started to ask about the possibility of the OECD getting it wrong.
Right away, I asked why we blindly accepted the OECD clearly being wrong about Canada’s fastest available internet speeds. To the OECD, Videotron’s world leading DOCSIS 3.0 deployment in 2008 just didn’t happen.
As it turns out, the OECD spends most of its time looking at incumbent telcos probably because most OECD countries don’t enjoy two separate wires into the household. As a result, the OECD frequently ignores the cable industry which, in Canada, has a 15% market share advantage over the so-called “incumbent” phone companies.
The most glaring error – and the lie that keeps getting repeated – was the OECD saying that on average, Canadians are paying $26.11 (USD) per megabit for our broadband. It isn’t true. It is just plain wrong.
How did the OECD mess up so badly on this one?
It turns out that the OECD collected its data by looking at advertisements – in the case of Canada, they looked at 16 ads: 5 DSL offers from Bell, 4 cable ads from Rogers, 4 cable ads from Shaw and 3 WiMax offers from Bell. There was no weighting based on market share; just price divided by Mbps. Add up the results, divide by the number of ads, there you go.
So Canada’s national average was skewed because the OECD happened to base 20% of its assessment on a product like WiMax which has less than a 5% share of the market. The OECD number was further skewed by giving equal weight to ads for Lite and Ultra-lite offers. So, by OECD logic of applying equal weight to each of the 16 ads, about half of Canadians are subscribing to wireless or light services – which have a much higher cost per megabit than the mainstream broadband services.
The OECD took the time to look at 71 ads from Australia, including 8 different offers from Bigpond at 8Mbps and 10 different offers from Bigpond at 30Mbps. There were 22 offers examined from Denmark; 36 from Finland; 36 from New Zealand. Denmark and Finland are countries with populations comparable to the Greater Toronto Area, yet the OECD did more than double the level of research for Finland than all of Canada. New Zealand has an even smaller population.
Australia has two thirds of Canada’s population, but the OECD looked at more offers from a single Australian ISP than it did for all of Canada. Had the OECD included multiple high speed offers from Canadian ISPs, its portrayal of Canada’s price performance would have been much lower. The OECD looked at more offers from a small Danish FTTH company with 6500 customers than it did from Bell Canada.
Why would the OECD use such poor data collection methodology?
Why are members of Canada’s media reluctant to critically look beyond the OECD headline numbers?
Why are Canadian government officials not getting such flaws fixed?
By the way, in the OECD analysis, there are only 4 WiMax prices examined in the whole study: 3 from Canada and one from Norway. The pricing from Canada was lower.
Despite the OECD fixation with some smaller FTTH plays, there was no mention of similar Canadian alternatives like Novus that offer fibre to the premises based services. Yesterday, Novus announced that it was upgrading each of its internet services by 10Mbps, with maximum data rates now going up to 60Mbps.
Update [August 20, 9:30 am]
Economist Suzanne Blackwell of Giganomics Consulting has done considerable work taking apart the OECD report and she has an excellent posting this morning that looks at Garbage In – Garbage Out that resulted in the flawed OECD results. She points out that the CRTC has found our price per megabit to be less than a quarter of what the OECD has reported, putting Canada into the ranks of the world’s least expensive broadband services.
That is a huge methodological error! How could the OECD have screwed this on up so badly? $26 per Mb? That doesn't even sound plausible.
They often do very good work in other areas. I'm disappointed this time.
Actinolite – You would have thought that the $26 per Mb would cause people to pause and think there must be a problem, but check out The Star and Citizen which cite the OECD study. Look at this CBC interview.
A flawed study with flawed results gets quoted over and over in the mainstream media – it must be true – and thereby takes on a life of its own.
Mark – asking why the media and federal government officials haven;t challenged the OECD's flawed methodology and data is valid. However, a more valid question would have been why the internet service providers themselves and their association, CAIP, haven't challenged the report's findings and presented the true story.