Cisco data about future Internet usage cited by NBN Co to justify its national broadband build fails to consider the impact of price, said Intelligent Business Research Services (IBRS) analyst, Guy Cranswick. The Cisco data showed a seven-fold increase in Internet data by 2016.
“The Cisco data is useful as a guide in that is shows how much traffic is growing and in what form, largely video,” Cranswick said. “But it does not show price factors on levels of usage.”
Older surveys showed high price elasticity on broadband and mobile, the analyst said. Rising costs on services like electricity and gas may also affect customers’ budget for broadband, he said. “NBN Co would have to demonstrate that data volume/usage is unaffected by price and they cannot,” he said.
“Our aim is to keep the prices we charge retail operators as low as possible,” an NBN Co spokeswoman said in response to Cransick's comments. “We recently committed to keeping the cost of our basic offerings frozen for five years and to limit future increases after that to less than the rate of inflation. Additionally, as we are more successful and see greater take-up of the National Broadband Network we have committed to further reduce prices.”
“This is because the NBN has not been established to maximise profits but to cover its costs, make an appropriate return and reduce wholesale prices over the long term,” the spokeswoman said.
Proposed pricing for the NBN should accelerate demand in the short term, but that may not continue over a long period, Cranswick said.
Optus and Telstra’s NBN price offers are “mid-high end,” not entry-level, he said. “That has a large effect as to general price for broadband as they have such large market share and they are characterised as serving middle Australia.”
To get a more accurate prediction of Internet usage, researchers should consider a range of groups such as early adopters, families and older users, Cranswick said. “Using IT journalists as a survey group would not be typical, and we’d have to have a large enough groups because extrapolating from just one or five instances is not strong enough to look at aggregate usage across larger groups.”
Cranswick said he’s “not aware of any [studies] of the scale required and over a long enough period of time.”
Follow Adam Bender on Twitter: @WatchAdam
Follow Computerworld Australia on Twitter: @ComputerworldAU