Thursday, October 13, 2005

ICT Reference Group Meeting

ABS House

10:30am 11 October 2005, at ABS House Canberra

Statistics aren't always dull. This week I attended a meeting at the Australian Bureau of Statistics at ABS House in Canberra:

"The ABS established an ICT reference group in early 2004 involving government, industry, academic and community representatives. The aim of the reference group is to improve the usefulness of ICT statistics in Australia from a variety of sources. The reference group provides a high level forum for understanding, improving and developing ICT statistics, providing members with the opportunity to discuss and consider strategies to address ICT statistical issues ..."
URL: http://www.abs.gov.au/websitedbs/D3110129.NSF/0/df543a3069a82fd2ca25702f0017e5e9?OpenDocument#Untitled%20Section_0

I was there as Director of the Communications Technologies (Telecommunications) Board of the The Australian Computer Society. The meeting was attended by about 20 people, one third from ABS, one third from other agencies and one third from non-government bodies, including AIIA, ACS, and Telstra. These are some notes I prepared (comments and corrections welcome).

ICT Satellite Accounts

A new macroeconomic report on ICT (to be called: "ICT Satellite Accounts") is due out in early February 2006:

An ICT satellite account defines ICT products and identifies their supply and use, so that a comprehensive set of economic data relating to ICT activity can be compiled for the Australian economy. Among other things, this allows us to quantify the size of ICT production relative to other types of economic activity.
URL: http://www.abs.gov.au/Ausstats/abs@.nsf/0/9ab6ab31699718f4ca256cdf00790444?OpenDocument

This should provide some impressive numbers for the media, of the “IT contributes $XXX billion to economy” sort.

Live demonstration - National Data Network

The National Data Network will provide infrastructure, protocols, standards, and services to support acquiring, sharing and integration of data across Australia.
URL: http://www.nationaldatanetwork.org/

The NDN looked very good. It is essentially a catalogue and gateway to Australian statistics at various organizations. There are some publicly available statistics there, but much of the data will require prior apporval to use, for priovacy reasons.

Information Development Plan

Information Development Plan (IDP) - The ABS is in the process of developing an IDP for ICT information. The ABS is taking a lead role in this development, but will not be the owners of the outputs of this process. There are many stakeholders involved in the production and use of ICT information and it is important for the success of the IDP process to involve and engage with these users and producers. The development of IDPs is seen by the ABS as an important element in progressing the National Statistical Service. Its effectiveness is manifested in how useful it is for decisions made on statistical priorities.
URL: http://www.abs.gov.au/websitedbs/D3110129.NSF/0/df543a3069a82fd2ca25702f0017e5e9?OpenDocument#Untitled%20Section_0

ABS doesn't have the resources to work on the IDP at the moment and will see if they can do it next year. The work DCITA has done on ICT productivity (reported at SEARCC 2005) could do with more support with more stats. James Shaw from DCITA talked about this at the ANU and again at SEARCC 2005 in September:

... we have the paradox that while a lot of computers and telecommunications are being used, economists are saying this don't increase productivity. ...
To explore this paradox, DCITA funded research to produce better measures of efficiency. These show about a 40 to 80% productivity boost with technology. Of course it is the Information Economy Division of the Department funding this work, so you might suspect the researchers are telling them what they want to hear. ;-)
URL: < http://www.tomw.net.au/blog/2005/09/ict-and-productivity-debate.html>

Broadband Statistics

The Australian Competition and Consumer Commission (ACCC) produces a regular broadband snapshot and is looking at changes:

The ACCC collects data on broadband take-up from a limited number of broadband providers, cross-classified by technology type and type of customer. This information is summarised in the ACCC's snapshot of broadband deployment reports.
The Minister for Communications, Information Technology and the Arts issued the Monitoring and Reporting on Competition in the Telecommunications Industry Determination 2003 (No. 1) (on the DCITA website) on 5 May 2003 under sub-ss. 151CMA(1) and (3) of the Trade Practices Act. The determination requires the ACCC to report quarterly at the aggregate level on wholesale and retail broadband availability and take-up, cross-classified by technology type, data speed, data usage, geographic postcode and business sector. ...
URL: <http://www.accc.gov.au/content/index.phtml/itemId/693170>

There was a long, and very familiar, discussion of what exactly is "broadband". ABS uses the definition of "always on 256k". This seemed to be the consensus and fitted with what I found when I asked ACS members about the government's Blackspots program. But in reality a claimed 256k connection might only give you 64kbps and even if you get 256k we should aim for more from a public policy point of view in the long term.

Others suggested a higher figure and more levels of measurement (ABS only measures up to 2Mbps). By the way my own submission to the previous Senate broadband inquiry was subtitled "Never mind the bandwidth, feel the quality". ;-)

One problem in conducting a survey of Internet access is that consumers may not understand what sort of broadband they have (whereas ISPs would). Another issue discussed was if wireless numbers should be collected. I suggested it was a good time to start collecting stats on wireless (have a wireless modem in my
Smart Apartment and noticed another one in the window of an apartment building opposite).

A fundamental problem with the definition of broadband is what are you measuring? As an example of the problems, one person at the meeting said they would compare dial-up use with broadband. But wireless ISPs now provide non-dialup non-broadband (less than 256kbps) services.

The ACCC is reviewing its broadband statistics for 2006. It is looking to collect more detail, such as: Retail V wholesale, postcode, transmission speed, business V residential. But this costs money and takes time. I suggested that the Internet itself might be used to collect statistics on subscribers and speeds. The routers in the network might be used to collect some statistics, for example. This got a sceptical reaction, but will be looked at. Might also be useful for DCITA's review of their Strategic Framework for the Information Economy (SFIE).

Voice Over Internet Protocol (VoIP)

There was a general discussion of what VoIP stats might be useful (and useful enough for people to pay for). It would cost money to include questions about VoIP in household surveys. I suggested asking the ISPs how much voice traffic there is on their networks. Some suggested ISPs have such stats but wouldn't provide them unless compelled to. The take-up rate of Digital TV then came up. I couldn't see how this fitted with VoIP, apart from being another gadget. Also I pointed out that having the gadgets didn't mean they were used. I bought a VoIP box and a HDTV card, but so far don't use them much.

Follow-up on Value of Information Propositions

Value of information propositions - This item was introduced at earlier reference groups, proposing guidelines for a process of determining the value of information. A more in-depth discussion took place, with the basic hypothesis being that information is valuable if it causes a decision change, and information affecting multiple decisions is as valuable as the most valuable use.
URL: < http://www.abs.gov.au/websitedbs/D3110129.NSF/0/df543a3069a82fd2ca25702f0017e5e9?OpenDocument#Untitled%20Section_0>

Couldn't understand what this was about.

Collection of Business Characteristics Data

Collection of business characteristics statistics - The ABS is in the process of conducting investigations into better integrating business characteristics statistics. It is considered that this integration will yield efficiencies and an increase in the usefulness of these statistics.
URL: < http://www.abs.gov.au/websitedbs/D3110129.NSF/0/df543a3069a82fd2ca25702f0017e5e9?OpenDocument#Untitled%20Section_0>

There was some discussion of how to increase the accuracy of the stats. I didn't know what they were collecting so couldn't comment.

Next meeting

The next meeting will be around March 2006. It will discuss Spam. So I wonder if ABS will e-mail a survey out to everyone about Spam? ;-)



No comments: