Re: Ratio of Lines to Subscribers (fwd)

[James Love <love@tap.org>]

  Pretty good set of data on the ratio.... and an interesting pricing
  system to deter nailed up lines (from his customers).  He gives the first
  120 hours free (about four hours per day), and then charges 20 cents per
  hour for the overage. (about 1/3 cents per minute).  He was a little 
  confused over what the PUC would do with this data (he thought we wanted 
  to regulate the number of modems for each ISP, rather that use the data 
  in looking at cost studies).  jamie
  
  ---------- Forwarded message ----------
  Date: Sat, 13 Jul 1996 14:16:11 -0400
  From: sechrest@jasmic.com
  To: Multiple recipients of list <com-priv@lists.psi.com>
  Subject: Re: Ratio of Lines to Subscribers 
  
  --------
  
  
  	The ratios of lines to subscribers are really dependent
  	on the behavior of the user base. And this depends on the
  	tasks, goals, and training the user base has. 
  
  	I have a unique experience with this, since I started
  	Internet services as a part of the Oregon State University
  	Computer Science Outreach Services. However, the University
  	decided this year that it really did not want to be in this
  	business and closed up shop. So I have moved those services
  	into the private sector. I don't know of anyone else
  	who has had the same business on both sides of the 
  	public/private barrier.
  
  	The clear issue is what level of business do the customers 
  	see. If I have a 10:1, or 8:1 or 1:1 ratio, these are
  	all the same if there are no busy signals. 
  
  	If you are optimizing costs, then you want to optimize 
  	so that you have just one modem that is never used. And
  	when that modem is used, you add more. 
  
  	But this puts things too close to the edge.
  	So people are generally more generous.
  
  	At OSU, we had times when we had modem ratios of 20:1 and
  	times when we had 10:1. Our experience was that 
  	a ration of 20:1 is unacceptable. There is far too much
  	access trouble. The level of major complaints escalate
  	about 18:1.  These numbers tended to have a 2 hour period from
  	8-10pm that were 100% full. 
  
  	At 15:1, we saw problems with peak performance every now and
  	then.... Like monday nights. But not very many complaints.
  
  	With a ratio of 10:1, we never had busy modems.
  
  	So.... While the milage may vary for people based on how thier
  	users are using it. We find that 15:1 is quite adequate. 
  
  	If you have a pricing scheme that balances the load, these
  	numbers will vary dramatically. IE. if my pricing is 
  	$10/month for 10 hours with $1/hour afterwards, 
  	then you will get a different user use profile than
  	if your pricing is 19.95 flat rate with no hourly charge.
  
  	(BTW, we charge $17.50 for 120 hours and $.20/hr afterwards)
  	This keeps the people how want to leave things on all the time
  	thinking a bit and hanging up the modem when they don't use it.)
  	
  	So, in effect, you don't care what the hours usages are, or
  	what the modem ratios are, since these are directly alterable
  	by your pricing policies. What you do care able is user service.
  	or how many busy signals that are being seen by the customer.
  
  	So it is not a good idea to use the modem ratio or user /hours
  	activities without the context of the pricing. And none of this
  	gives you a measure of system performance as seen by the 
  	user, except indirectly... 
  
  	So... I would recommend that the PUC not be involved in any
  	setting of regulations based on modem ratios or hours. It will
  	just create artificial structures that do nothing to improve
  	customer services.
  
  
  ------
  John Sechrest         . 
                         .                      
                           .        Public Electronic Access to Knowlege
                              .                 
                                  .              Corvallis Oregon
                                        .                 
                                              .          sechrest@peak.org
                                                        .
   Helping people use computers and Internet more effectively        .