T H E 
B O T T O M L I N E
Too Big for Your Own Good
The demand for Internet access is so strong that many companies are unable to provide adequate service. Learn how to get the best deal.
by Joel Snyder
The internet was the number-one growth business in the first half of this decade, and may continue in that position through the end of the century. Unfortunately, it is no more immune from growing pains than any other business sector. Among the casualties of the Net's skyrocketing growth are users who are receiving poor service from Internet access providers--large and small alike--that cannot keep up with overwhelming demand.
Among the spectrum of Internet service providers selling access to businesses and individuals are large firms such as BBN Planet, CompuServe, Netcom, PSI, Sprint, and UUnet as well as small local providers like the guy down the street with eight modems and a 56-kbps line. Many of these mom-and-pop operations have sprung up, eager to cash in on the Internet.
The problem is that all of these companies do not seem to understand the word "service." Service providers have to do more than promise you a piece of the Internet. They have to be there to run their business so their customers can run theirs. When something is broken, they have to fix it. When you call, they have to answer the phone. As the businesses age, they have to plan for growth so everything doesn't collapse around them and their customers. That's the other half of a service business, and that's the half that's missing or deficient throughout most of the Internet.
All of the service providers I've dealt with (and as a consultant I deal with a lot of them) are now operating under a state of siege. They have double- and triple-digit growth rates and just can't handle it. Managing growth is a classic problem in any business. It just happens to be hitting everyone providing Internet services right now.
Hope and Pray Approach
One of the reasons service is so bad is that the equipment generally is pretty good. Unlike the computers of yesteryear, you can stash an Internet service business in any small closet and, if things go well, not have to touch it. The Internet business is a 24-hour business. It's expensive to employ people in a network control center when 99 percent of their time is spent twiddling their thumbs. That means it's very tempting to not have a network control center at all--a temptation to which many Internet services businesses succumb.
Of course, if Unix is involved it's an entirely different matter because Unix requires constant preening and watching to make sure it doesn't make expensive messes all over its disk drives or shut down because a bit got twiddled in the wrong place by a passing gamma ray. But even the presence of Unix doesn't seem to inspire the 24-hour coverage such a business needs.
A second reason for poor support is the price wars in the Internet business. Most service providers jump into the business without any real business sense or realistic plans. They enter a market where the going rate for service is so low they feel the only way to gain market share is to underprice the competition. The prices inch down, dollar by dollar, until there's no money to pay for good round-the-clock support people, for equipment expansion and replacement, and to make sure there are no busy signals when users dial in.
Statistics vary from observer to observer, but an astonishingly large number of businesses--something like 80 percent--never make it past the growth stage. They falter, for one of a hundred reasons, and are never heard from again.
A few businesses have enough money to wait out their under-capitalized competitors. When the competing firms finally fail, prices go back up to rational rates and things settle down--until the next entrepreneur comes to town with $50,000 to waste. I just created a business plan for an Internet service provider who wanted to sell dial-in SLIP and PPP service. We calculated that the company would have to attract almost 10,000 subscribers to make a reasonable return on investment if it sold accounts for $20 per month.
Worth the Price
Look at the big information service providers such as AOL and CompuServe. They charge a hefty hourly rate because that's what makes economic sense. They've got all the advantages of scale that no smaller providers have, yet they can't afford to offer the service at the same price as Bargain Basement Net down the road. If you're paying $20 per month or less for dial-in service to the Internet, your provider is either hemorrhaging money or seriously shortchanging you on the quality of the service. I challenge anyone reading this to send in a business plan that shows how you can make money and provide good service at those rates.
A third reason Internet providers are providing such lousy service is that they're so dependent on other businesses to support them. For example, if I need more phone lines for my business, I have to wait between six and eight weeks to get them. That's an eternity by the Internet clock. If I want a particular brand of modem, I might not be able to get it quickly because some modem manufacturers have not been able to keep up with demand. The same is true for other kinds of equipment, such as communications servers and routers. And if I should need another high-speed line to the Internet, forget it. I have to buy equipment from two different vendors, service from three more companies, and schedule time with two different consultants to get it all installed.
One of the service providers I used to deal with is a pioneer in the field and has been in business for many years. However, this company has become so sloppy about its service and network that it's not worth doing business with the firm anymore. As I was canceling my account, the provider was 90 days into a network problem that it hadn't fixed. The provider blamed the phone company, but I blame the provider for not finding a way around the phone company. While one week is reasonable, three months is not. My best guess as to why the company hasn't taken other measures is that it would cost too much money.
What You Can Do
All of these factors have combined to ensure that it's virtually impossible to get good Internet service in some cities. Now that you know why you've been getting such bad service, what can you do about it? Here are four solutions.
First, make sure you know from whom you're buying service. If you're like me, you're not interested in jumping from company to company, even if it's only for basic dial-in service. This especially applies to business connections. To switch providers, you might have to renumber your entire TCP/IP network, and you don't want to do that if at all possible. So take some time to investigate the company. How large is the company and how long has it been in business? Do the customer service reps make good business sense when you talk to them, or is the deal you're getting too good to be true?
Second, find out as much as you can about the company's network. Any Internet access provider should be able to draw you a physical layout of its network, showing the links all the way back to one of the NAPs (network access points), which are the core of the Internet backbone. You may be going through several companies to get there, but your provider sure better know the route your packets are going to take.
What is the ratio of bandwidth sold to bandwidth going towards the Internet backbone? If the service provider has a single T1 (1.54-Mbps) line to the Internet, it should have no more than 100 modems for dial-in customers, and preferably 50 or 75. If the provider is selling 56-kbps services, the ratio should be even lower--about 25 or 30 such lines per T1. If your provider has more than one T1, find out whether you are benefiting or not. It may be that all your traffic has to go over one line, while the other circuits are used for other locations or different types of customers.
If you're buying dial-in service, find out the ratio of modems to customers. Five-to-one is ideal and probably guarantees you no busy signals. Ten-to-one is what you'll more likely encounter. Any more than that and you're not going to get in when you want to.
Third, test the support and facilities yourself. If the service provider claims to have 24-hour support, call 10 or 15 times during a week at all hours of the day and night. Is someone there to answer your questions or do you get shuttled into voice-mail limbo? Dial in many times at all hours. Are you getting busy signals or no answer? If so, there's a problem.
Fourth, get your terms in writing. We like to trust people, but if your business is on the line, you need more than the promises of salespeople. If they say their service doesn't have busy signals, insist on getting a month's credit if you get one. If they say they never go down, insist on penalties when they do. Honest and reliable businesses are happy to put their promises in writing.
Also make sure the rate you're paying is reasonable. Don't go for the lowest bidder because the firm probably won't be around long enough to save you money. If you want a better deal, offer to sign a year contract to lock in a better rate. You may not get perfect service, but if you do your homework, you can get the best service in town at reasonable rates. 
Joel Snyder (jms@opus1.com) is a senior partner at Opus One in Tucson, Ariz.
Reprinted from Internet World magazine Vol. 6 No. 12, (c) 1995 Mecklermedia Corporation. All rights reserved.
