Who will win the "last mile" market for Internet Service?

Consider the "war" between the telephone companies and the cable TV companies over who gets to deliver electronic services to your house. To some degree, I believe that the "war" is a fiction, because who says I can't have more than one wire or fibre into my house? I do now: telephone, cable TV, and even power from my local utility company.

However, when it comes to delivering a data service, I believe that the telephone companies will be the ones to offer the best service over their infrastructure. There is a perception that there is more bandwidth available for data in the typical cable TV system, but I believe that most such systems are unsuitable for data.

The principle problem is a matter of signal characteristics, and the economics of the two industries. Television is a "loss-tolerant" signal; i.e. a little fuzz in the signal is not likely to be noticed by the typical customer, precisely because there is so much signal. What's a dozen lost pixels in a typical TV frame going to mean to the overall picture quality? Not much difference.

I claim the effect of this is that most cable TV companies have wired their "cable plant" with cheap cable, and cheap equipment. After all, if a little signal loss isn't going to be noticed, why bother to even try for 100% signal fidelity?

Also, television isn't taken as seriously as telephone. How often does your cable TV service fail? If it does fail, do people potentially die from it not working? By contrast, if the telephones aren't working, it's a really big deal - everyone gets upset, and the telephone companies take their responsibility to deliver a reliable serivce very seriously. This is why they can and do give service guarantees (99.4% availability, or some such).

In the end, I believe that the market for Internet service to the home is the telephone companies' market to win, or to lose, depending upon how they play it. They have the experience in delivering reliable point-to-point data services with guaranteed levels of performance, and large customer service organizations to back them up. Their major issue is that for any data service more complicated than a point-to-point bit pipe, they've generally screwed it up. Whether they will learn to do the Internet right is anyone's guess.

The Cable TV companies have bad plant, and poor reputation for both reliability and service. With the Direct Broadcast Satellite Television companies (e.g. DirecTV, PrimeStar, EchoStar) putting serious competitive pressure on the cable TV companies, what is the likelyhood that they will upgrade their bad installed plant well enough to do real data services?

As the old joke goes, would you buy a used bit pipe from this man?


Erik E. Fair <fair@clock.org>
April 26, 1996