Telecom Product

Telecom, of course, is short for telecommunications. Telecom products fall into two main categories: commercial and consumer. For now we will concentrate on the consumer telecom market.

The telecom market, including the consumer telecom market, can fairly be described as "complicated." The pubwan movement is committed to a NeutralityPrinciple, so we will avoid controversial questions like whether it's the most complicated industry in existence, or whether it's complicated by design. That it's complicated is hopefully uncontroversial enough to merit a discussion here.

One of the complexities facing the consumer of telecom product is the presence of "dilemmas":

1. mobile or "land line"
2. pre-paid or "contract"
3. listed or "unlisted" (note that this one reverses signum when one transits between

   the residential (individual?) and business (institutional?) environments.)

   bundled or unbundled (this is not really a di-lemma but a multilemma because each bundle contains its own product inventory).

   buy or lease the actual tangible product, i.e. the phone

1. which telecom product bundles
2. offer which features
3. with which strings attached
4. at which prices
   This should include the capability to monitor all these variables simultaneously and in "real time" (i.e. using information that isn't too stale). The pubwan database should also retain "historical" information. The coding of this information as "data points" is covered in the essay Pubwan Virtual Objects One of the first steps in modeling PVO's is identifying "atomic goods." By this we mean the "line items" on an inventory of goods offered for sale as a bundle. Like any exercise in data or object modeling, this is an inexact science. Also, like any exercise in consumer (self-)education, it is also an exercise in advanced disinformation theory. The present author advises trying to think of the whole process as a game. When trying to decide whether I have reached the "atomic" (SPVO) level of "analysis," I consider certain questions. Basically, I try to make a reasonable assumption about whether a reasonable consumer would think of two goods as two brands of the same commodity (or "gendesc"), or as two different commodities. To one consumer, Euvax brand pickled eggs may be substitutable for Javonip brand pickled eggs, while for another they are not. Likewise, another consumer may insist on "all natural" products and reject pink (food dye) pickled eggs, regardless of brand or supplier. This is enough getting sidetracked on pickled eggs in an essay about telecom product. We were discussing the question of what information is contained in a data point. The pickled eggs examples are to demonstrate the fact that which descriptors are "essentially" generic descriptors (gendesc) is SUBJECTIVE. For this reason, where to "prune the tree" of descriptors for the purpose of comparing bundles should be customizable for each decision support client. On the other, hand, information about a participant's personal sense of what level of product differentiation implies non-substitutability is best made public. The creation of a descriptor tree comprehensive enough to be useful cannot be done by a small number of participants. This is where the concept of critical mass comes in.

   After the process of deconstructing (actual) bundles comes the process of constructing (virtual) market baskets. First, there we construct the set of "atomized" market baskets. Fortunately (for this purpose if no other) this is constrained by income so it is FINITE. It is also constrained by pricing policies, so it is also NONLINEAR. This is the part of the exercise that demonstrates entropy by demonstrating that information is ridiculously easy to destroy and ridiculously hard when it comes to "putting Humpty together again." Then, we determine the nondominated subset of this set of atomized market baskets.

   Let's assume for the sake of argument that we have established a list of available and affordable bundles of telecom product. Remember that the parameters listed below refer to properties of bundles, not of their constituents.

   1. total up-front price in $ (min)
   2. total marginal spending rate in $/sec (min)

      (This is a fancy term for "monthly subscription rate." We tabulate it by the second instead of by the month, because the second is the SI unit of time, and the month isn't even plainly and unambiguously defined.)

      percent of time spendable online (dimensionless) (max) In theory, having "unlimited" internet connect time means the ISP doesn't mind if you log in and never log out, or 100% of the time online is OK. In practice, there's some kind of contractual fine print that says the ISP doesn't hold the user to a maximum number of hours, but also that there is some kind of "no free parking" clause. This, of course, is especially explicit when the ISP is a subsidiary of the phone company :) "30 hours a month" means we divide 30 by the number of hours in a month (on average, 730.45) and use this percentage (0.041=4.1%) as the value of this parameter.

      percent of time spendable on phone (also dimless) (max)

      total number of email addresses (maxlo) Notice that so far I have suggested signa (minimize or maximize) but not priority (high or low). This is because I hypothesize (for admittedly intuitive reasons) that signum (desirable vs. undesirable qualities) is fairly "objective" while priority is very "subjective." I figure just about anyone would maximize their number of email addresses (because two units of any good is better than one) but with low priority. The main reason I make this assumption is because the number of possible "well-formed" email addresses is humungous, compared to, say, the number of possible well-formed phone numbers.

      total number of phone numbers (max)

      total number of mobile phones (max)

      total number of pagers (max)

      minimum number of pager characters (max)

      average number of pager characters (max)

      espn3 included (maxlo)

      static IP included (maxhi) See SamIzdat. Some people prefer the stick over the automatic.

      late payment fee (min) Minimize this with high priority if you consider yourself a "bad credit risk," or to make a statement against in loco parentis. I call this "de facto interest." Legally, it isn't interest, but its IRR as a cash flow can be in triple or even quadruple percentages! Caveat emptor!

      works with Linux (max)

      works with nonproprietary network interface (maxhi)

   When listing objectives for the purpose of multi-objective optimization, the possibilities are literally endless. Keeping these lists short is generally a good idea. Long lists of objectives result in the total absence of "domination" in the Paretian sense, which means your efficiency frontier will be all inclusive and therefore contain no information. The idea is for the pubwan database to contain a huge list of objectives for any conceivable product class, but for individual queries against the pubwan database to be concisely specified.

   ---

   Updated: 2004-07-09

   Home / Index

   Hosted by www.Geocities.ws