We reasoned at that time - and still do - that a simple count of the number of registered domains is an inadequate way to measure the true number of Australian web sites, for the following reasons:

• Not every registered domain has a web site attached to it. Many companies register a domain purely for email purposes. Many other domains have been registered by companies who plan to use them at some future time, but have yet to do so.
   
• Many web sites - especially in 1996 and 1997 - exist under subdomains or on subdirectories of their ISP's or host's web servers. In the earliest days of the Net in Australia, domains weren't seen as particularly important by many legitimate site operators.
   
• Not every Australian web site has an .AU domain (eg: Telstra.com, BigPond.com)
   
• Many domains are multiple registrations owned by the same company which point to the same web site (eg: a manufacturer may register a domain for each of its major brand names or trademarks; a newspaper proprietor may register a domain for each publication in its stable) 
    
• Many domains have been registered for purely speculative purposes. Their owners have no intention of ever attaching a web site to them and are simply hoping to be bought out (eg: aol.com.au was originally registered by a speculator, and was subsequently acquired by AOL when they opened operations in Australia).
   
• Because most .AU domain registrations last for two years and the failure rate of web site operators is notoriously high, many "dead" domains exist in the official registry at any time. They aren't weeded out of the registry until the owners fail to renew them.

Instead, we reasoned that any legitimate and serious web site would list itself in the major search engines and would include at least some geographic information in its text (eg: the site operator's office address).

Furthermore - since the overwhelming majority of Australia's population live in the country's capital cities - we reasoned that by searching spider-driven engines for these capital cities we could derive a rough approximate of the actual number of truly "live" Australian sites, regardless of what domain they operated under.

We recognise that this is a far from perfect methodology. But in essence, it's very similar to the well-known ANZ Jobs Index (which attempts to estimate trends in employment by counting the number of job advertisements that appear in capital city newspapers).

This may also explain why the estimated number of live sites we report is only about 33% to 40% of the number of domain registrations that exist at any time.
    

Changes In The Index

The reason for this is that although we always poll multiple engines to derive an overall average figure for each capital city, we found that changes in the way the engines operated during this period necessitated dumping some and moving to others as time went by.

One serious issue that emerged during 2000 was that we noticed that many former "name" search engines appear to have simply stopped adding new sites to their databases.

Our monthly polling clearly showed us that these engines were returning the same numbers to our queries from one month to the next. Our own site statistics confirmed that these engines were no longer sending out spiders for indexing purposes.

This discovery tends to confirm a widely-held suspicion amongst some Internet marketers that many of the earliest Net search engines have either "maxed out" their capacity and/or have effectively been abandoned by their operators, though they still give all outward appearances of being live.

This discovery has important implications for both web site owners and search engine users.

• For web site owners, it may indicate that their ability to gain exposure in some very well-known search engines is now compromised (ie the engines still accept listings, but do not actually add those listings to their files).
   
• For search engine users, it may indicate that some major search engines are now seriously out of date and are becoming increasingly irrelevant as true guides to the Net from one month to the next.

Naturally, we welcome any correspondence on the AIGI - though we hope this answers the most common questions you might have about it.