[1971] in Commercialization & Privatization of the Internet
re: ANS Pricing Model
daemon@ATHENA.MIT.EDU (Guy Almes)
Tue Jan 14 08:22:47 1992
Date: Tue, 14 Jan 92 8:16:39 EST
From: Guy Almes <almes@ans.net>
To: Jeff.Bone@ebay.sun.com (Jeff Bone)
Cc: almes@ans.net, com-priv@psi.com
> From: Jeff.Bone@ebay.sun.com (Jeff Bone)
> To: com-priv@psi.com
> Subject: ANS Pricing Model
>
> I have a few questions and comments on the ANS Plan.. If these are naive
> or ignorant, please be patient. Standard disclaimers, of course.
Jeff,
I'll take a hand at it. Let me know if this addresses your issues
appropriately.
-- Guy
> I don't understand the reasoning behind the "attachment fees" paid to ANS
> by the midlevels for end customers which attach to the midlevel. Why in
> the world does ANS get a piece of every single customer of every midlevel?
> It would seem to be more reasonable for the midlevel to pay only for the
> size of its data pipe(s) to ANS (and its COMBit traffic), and have what
> happens on the midlevel side of the gateway be essentially invisible to ANS
> from a business/revenue perspective.
There are two answers:
<1> This aspect of the Plan is open to negotiation; we're generally receptive
to ideas from the mid-levels. The Plan is *not* a 'take it or leave it'
document, but rather our initial cut at what we think will be of mutual
benefit.
<2> Now, as for why it's part of our initial cut: we want something that is
affordable for small mid-levels and scales well for large ones. Our current
open suggestion is to use a schedule that gives us less revenue for the pure
data pipe, but has the revenue grow slowly with the growth of the mid-level.
> Allowing ANS to bill the midlevel for each customer that midlevel attaches
> seems to have some frightening implications:
>
> (1) end customer attachment fees are anticompetitive
>
> By "fixing" a category/price structure for attachment, ANS in effect gains
> the ability to control the baseline price for connectivity. The midlevel
> must pass the attachment fee along to its customers in order to recover
> cost, therefore ANS sets minimum prices for everybody. This reduces the
> ability of the midlevels to compete on a price basis. The effect of this
> might be minimized if the midlevels were given the ability to participate
> in determining the attachment category/price structure.
We have no motivation to make the mid-levels uncompetitive. If they're
uncompetitive, then our success is reduced. In fact, let me turn it around:
if we had no incremental revenue with the growth of the mid-level, we'd have
to charge more for basic connectivity. That might be acceptable to the very
largest mid-levels, but it would prevent the smaller mid-levels from
effectively entering the market. *That* would be more likely to be
anticompetitive. It is in *that* case that we'd "control the baseline price".
Maybe the difference in our intuitions on this matter is that you imagine
the incremental revenue per new mid-level customer being large: large enough,
in fact, to make it hard for the mid-levels to sign up new customers or large
enough to make it financially uninteresting for them to do so. That situation
would hurt us, and we'll work with mid-levels to avoid such a situation.
> (2) inhibits "very low end" connectivity
>
> The attachment fees listed in the document would tend to indicate that
> connectivity is still going to cost the end user a good bit. What does
> this do to the ubiquitous networking idea? One goal (at least in my mind)
> of the whole commercialization effort is to make the resources more widely
> available. Unless very low end service can be priced at a level comparable
> to phone service, the low end will not flourish. Given the ANS plan, it seems
> to me that it would be very difficult for midlevels to profit by providing
> very low end service.
I take it that by "very low end", you mean for low-speed connectivity to small
organizations. In such cases, the fees are either very small or zero.
> (3) gives ANS unnecessary leverage
>
> The costs-to-operate which ANS incurs on a per-midlevel basis are very
> closely related to the bandwidth of the connection and the overall
> traffic and usage of that bandwidth. Getting a chunk of change for
> every customer on the other size of the midlevel's gateway is great for
> ANS, but is IMHO greedy and unnecessary. In addition, ANS apparently
> intends to compete with the midlevels for end customers (I gather this
> from the Cooperative Agreement statement). SO --- ANS gets the buck
> whether or not they get the customer. I'm sorry, but that's not the
> way a free and open market works!
Two points:
<a> With regard to "greedy and unnecessary": the key virtue of incremental
pricing per customer is that it allows the our data-pipe charges to be lower
than they would otherwise be. Our intent is to allow mid-levels to be able
to start small and grow.
<b> With regard to "ANS gets the buck whether or not they get the customer":
the incremental fees are small enough that maybe we get a buck if we connect
to the customer directly or make 15 cents if the mid-level connects them.
> My suggestion: charge for data pipe capacity only. If the midlevel
> signs more customers than can be accomodated by the single pipe, ANS
> sells them another pipe. This seems fair and equitable.
>
>
> Okay, somebody clue me in: why are these arguments ill-formed and
> wrong? :)
There's nothing wrong or illogical with your questions or ideas. I think the
gap between your stance and our plan rests on two things:
-- you might not appreciate the extent to which the increment per-customer
fees allow us to lower the data pipe capacity fees. This is an issue where
big mid-levels and small ones will naturally have different views. Our
plan was developed in close consultation with several mid-levels, both big
and small.
-- you might not appreciate our motivation to make attachment of new customers
by the mid-levels a win for the mid-level, the customer, *and* for us. Our
emphasis is on high-speed attachments. A heathy mid-level is a natural
user of such attachments. Most direct customers are not.
Jeff, I think your ideas are constructive in themselves but may reflect an
imperfect understanding of our stance. I hope this note can contribute to
resolving this.
Cheers,
-- Guy
> Yrs,
>
>
> Jeff Bone
