John Linton
The end of the latest financial year is less than three weeks away and I have no recollection of where it's gone. I think that's partly because I have spent most the the past 11 plus months worrying about how the threatening financial situation around the world and especially in Australia would affect Exetel's ability to survive and if we were able to ensure it survived then what needed to be done to ensure it was able to grow and make the most of whatever opportunities came our way. As it turns out I seem to have wasted a whole year worrying about scenarios that never materialised and the twelve month period has turned out (given nothing changes in the next 19 days) pretty much like every other year of Exetel's existence.
There is one major difference though. That is that the average data usage by residential customers has increased more rapidly over the past 12 months than in any 12 month period I have records for which is the past seven years. I was talking to a nice industry analyst today and I remembered that I have been heavily involved in the provision of internet services since the mid 1990s when 14.4kbps dial up was the only widely available residential service with 28.8 kbps still many months away. No need to concern yourself with data usage by customer in those days.
I do understand that Exetel's 'average customer' is very different from Bigpond's 'average customer' and also different to almost every other ISP's 'average customer'. So what their average download by customer might be will be different because of the huge differences in the demographics of each company's user bases. My involvement with broad band goes back to 2002 and my records show the 'average usage over the years increasing rapidly but not anything like as rapidly as over the past year or so. The average monthly download per customer for the last few years looks something like this:
2002 - 1.1 gb
2003 - 1.6 gb
2004 - 2.0 gb
2005 - 2.7 gb
2006 - 2.9 gb
2007 - 3.3 gb
2008 - 3.6 gb
2009 - 4.5 gb
A 25% increase over a 12 month period is a very, very large growth and if it holds true for the coming 12 months it will pose an interesting problem. Pricing of 'pure' IP bandwidth has fallen by about 30% in 2008 and by a further 2% or so in 2009 so pricing to the end user can be accommodated by Exetel in IP terms but it is the bandwidth between the customer and the Exetel switches that will present the problem in cost terms because it hasn't fallen at all so far in FY2009 though there were some significant reductions in FY2008 for ADSL2 connections. You also have to remember that Pipe peering, Akamai and PeerApp caching all combine to reduce the 'pure' IP pricing component but there is no such amelioration on the 'other side' of the network.
As far as I can see, at the moment at least, the 'pure' IP cost reductions will reduce the overall cost of the data supply cost of the service based on today's significantly increased 'average' per user usage. However if the average downloads per customer keep increasing at the same rate as over the previous 12 months they won't continue to do that. It's an interesting 'problem' because it can be solved in only one realistic way (I am discounting raising the plan costs as inappropriate and quite hard to do) - by adding new customers whose average usage is much less than the average usage of current customers.
Problem solved....or it would be if I knew how to do it - which I don't. At the moment our monthly ADSL order intake is continuing to slowly increase each month and I really don't want to play around with that scenario. If the advertising we do to promote the HSPA service is 'successful' it will be of assistance in reducing the average download per customer but it's unlikely to have much effect over the next few months. So, basically, the only option is to change/move the target demographic for the ADSL plans in to 'unknown' territory.
While I'm not really a timorous person (we opened a company in Sri Lanka and bought our own floor space in what were meant to be difficult financial times this year) I really have been dithering over what to do about 'finding a new broad band market' for Exetel. I have prepared a new set of plans for both ADSL1 and ADSL2 which make financial sense but they don't appeal to me 'aesthetically' as they move away from the direction we have been heading in for the past 5 plus years which appeared, at least to me, to be finally 'paying off'.
It's almost always dangerous moving target demographics especially when you have spent many years understanding one marketplace and have paid little attention to any other. I was surprised to be told today that 'general industry sources' put the annual churn rate of ADSL users at something like 25% and that the reason for churn was people were looking for a better service experience (as in help desk type service). While I can understand that concept I don't see it in the figures from our data base which puts our annual 'churn away' rate as something less than 7% with 'churn ins' at around 22% - so that information has further confused me.
In our, absolutely not meaningful, small numbers it appears to me that people choose Exetel because of a lower plan/speed price and larger than 'average' download allowances. While far more people churn to Exetel rather than away from Exetel that is, mainly, simply a function of Exetel's tiny market share of users available to churn away versus the 50 times larger base of users available to churn to us. Yes. I know, statistics say that shouldn't be true but it is in this case.
So if we throw away the base premises we have been refining for over five years what size of risk are we taking?
I guess there's still a couple of weeks in which to make a 'final' decision.