The San Jose Mercury News


PITFALLS OF AN INTERNET PROVIDER
WHAT A DEAL, SIMSON GARFINKEL THOUGHT. HE'D START A CO-OP AND GET CHEAP INTERNET ACCESS FOR HIMSELF AND HIS FRIENDS. SUCH A DEAL


Published: Monday, September 2, 1996

Section: Business Monday

Page: 1E


By SIMSON L. GARFINKEL, Special to the Mercury News

NO MATTER how we did the arithmetic, the numbers didn't add up.

Each time my business partner Eric Bates and I signed up a new customer for Vineyard.NET, our Internet Service Provider on Martha's Vineyard, we lost money. Even with the $30 sign-up fee and our $25-a-month membership fee, we were losing $75 to $100. We weren't breaking even on our new customers until after they had been on the service for three or four months.

There was only one thing to do: raise our rates.

These days, it's nearly impossible to find an ISP that's raising its prices. With the cut-throat competition and new providers like AT&T and MCI springing up every month offering ''unlimited Internet access'' for $19.95 a month or less, a price-increase should have been the last thing on our minds. But we've done a lot of unconventional things with Vineyard.NET, which is probably one of the reasons why it is still in business, and why I haven't gone insane.

For the past seven years, I've been a free-lance journalist who writes about science, technology, computers and the Internet. A little less than two years ago, my wife and I were enjoying Christmas on the island of Martha's Vineyard, seven miles off the Massachusetts coast. On that fateful vacation, we made the mistake of looking in a real estate guide and discovered that we could purchase a four-bedroom, 2,000-square foot house on the Vineyard for less than $170,000. We started looking at houses that spring, and purchased our own 150-year-old Cape Cod a couple of months later.

There was just one problem: I needed a full-time Internet connection to do my job, and there wasn't an Internet provider on Martha's Vineyard that was up to the job. A leased-line to an Internet pro vider in Cambridge would cost $450 a month. Rather than simply paying this money out of my pocket as a legitimate business expense, I decided to start up a small, ''cooperative'' Internet service provider and split the cost with any other islanders who joined.

In retrospect, it's too bad I didn't stick with that plan.

The first two people who signed up for the cooperative, an electrician named Bill Bennett and Eric, a local carpenter, decided that they wanted to turn the cooperative into a business. Why not? Internet companies were going public every week, making hundreds of millions of dollars.

ISPs in Boston were doubling in size every six months. None of us had any idea about the actual economics of running an ISP, but that didn't matter. Didn't all ISPs charge more or less the same rates and use more or less the same equipment? Surely all of those other companies wouldn't be in business if they weren't making any money.

(Being a journalist, I was also concerned about the potential conflict-of-interest in running an Internet provider and writing about the Internet. I solved that problem by disclosing my relationship with Vineyard.NET whenever I felt it was appropriate to do so, and by trying not to write about direct competitors.)

One of the things that was surprising was how much effort was involved in setting up an Internet provider that had nothing to do with the Internet. We easily spent 100 hours getting incorporated, getting health insurance, setting up our bank account, getting approved to accept credit cards, getting checks printed, and more.

By September, we were signing up a dozen people every week. It didn't take long before the phone calls started - usually during dinner. Rae couldn't log in: she thought that something was wrong with our system. Myles couldn't log in: he had accidentally re-formatted his computer, and needed the Internet software installed again. John wanted to know why he wasn't getting e-mail. James wanted to know how to download the pornographic photos.

My telephone was ringing day and night. You see, Vineyard.NET was located in my basement. Bill called it the ''brain room.'' We had to bring in three electrical circuits, pour a concrete slab, and build a wall for security. Somehow we convinced the phone company to pull a telephone cable with 100 pairs of wires from six blocks up the street to my house. We also got an uninterruptable power supply with four hours of battery backup (provided by two automobile batteries) because electricity on Martha's Vineyard is notoriously unreliable.

Anyway, back to the ringing telephone. I had given my phone number to the first few customers while we had waited for a business line to be installed. Our local newspaper compounded the mistake, advising people to call my home phone number for further information.

Pitfalls: costly tech support, clueless customers

I didn't know it at the time, but it turns out that technical support is one of the most expensive costs in running an ISP. Networking software is fairly difficult to install, and every PC and Macintosh is a little different. Compounding the problem are folks who have old computers that aren't really capable of running the new Internet software. And then there are the customers who are simply clueless.

We soon discovered that it was taking between 30 minutes and an hour to set up each new customer. But we couldn't charge customers if it took longer than that - customers thought that they were purchasing a service from us, and it was our fault if it didn't work, not theirs.

And if they had any other problems with their computer, or their software, or the Internet in general, they called us. Somehow, we had implicitly given our customers an offer of unlimited technical support. America Online, Netcom, and practically every other on-line provider has this same problem. That's why you have to wait 30 minutes on hold to ask questions to a live person: It weeds out the bored users who are just looking for someone to talk to.

I spent 40 hours a week for the first few months working on Vineyard.NET (remember, this is not my principal vocation), when I was writing the accounting software. Now I'm down to roughly one hour a day. But it's a bottomless pit: I can spend as much time on Vineyard.NET as I want. There's always more to do.

It's at points like this that many ISPs make their deal with the devil, and step on the treadmill of unbounded growth. If Vineyard.NET had been my career, rather than a way for me to pay for my own connection, I would have tried to sign up as many customers as possible, used the quick infusion of sign-up fees to hire somebody to help me with the technical support, then started advertising to attract even more new customers.

A number of factors kept Vineyard.NET from going in that direction. First, there was the lack of money. Then there was the geographical problem: our service was limited to the island of Martha's Vineyard, which has a year-round population of 12,000 people. We simply didn't know how far we could expand before we would have to start going off-is land to get new customers.

Personal concern: keeping
writing career going

Lastly, there was the personal problem: any aggressive expansion of Vineyard.NET would have meant the end of my writing career. Frankly, being a journalist is a lot more fun than running an ISP.

Our solution was to partner with the Island's computer stores. I wrote a series of programs for the Vineyard.NET web server which let employees at the computer stores set up accounts on the Vineyard.NET system. Then I wrote an accounting package that kept track of how much money each customer owed us, and credited the stores 20 percent of the customer's monthly fee when we received payment.

Since the computer stores weren't part of Vineyard.NET, they could get away with charging customers for extra technical support. They could sell customers new software or hardware that their computer systems might need to work on the Internet. And they could use the customer contact provided by Vineyard.NET as an opportunity for selling unrelated products - something that we couldn't do because it would have looked bad.

Of course, I had to write a lot of programs to create this system for letting the computer stores manage our accounts. The entire system management and accounts receivable system, complete with credit-card billing, runs to nearly 20,000 lines of computer code, all written by Eric and myself over a six-month period. But the thing is, even if we hadn't joined up with the computer stores, we would still have had to write this accounting system from scratch.

That's because there was no off-the-shelf accounting software that worked on the Web. If you wanted to send out your invoices via e-mail and let your customers view their account balance with a Web browser, the only choice was to write the software yourself.

(Since them, I'm told, a few companies have started selling such commercial software. But I have yet to discover an ISP that is actually using a program that they have purchased, rather than writing their own. That's because this kind of software is deceptively simple to write. Indeed, I had the basic system working in less than a week. The rest of the time was spent debugging and adding features.)

Finances: Profit equals income minus cost

This September marks Vineyard.NET's first birthday as a commercial Internet provider.

Are we making any money? It's hard to say. Bill recently left the company because he thought there was no way that Vineyard.NET would ever make a decent profit. I'm inclined to agree.

When I took my course in micro-economics, I learned that profit equals income minus cost. The income of an ISP depends entirely on how many customers you have and what they are charged. The cost is determined by your system's capacity.

Like all ISPs, Vineyard.NET has a limited number of phone lines - considerably fewer than its total number of customers. When a customer wants to use the Internet, they call up one of our phone lines, initiate a connection with one of our modems, and start transferring data to and from the Internet over our high-speed connection. The hardware necessary to service one customer is called a port. We can get by with having fewer ports than customers because not every customer calls up at the same time.

Equipment: using modems stacked 16 to a rack

Ports are not cheap. The phone company charges us $50 to install each new phone line. Off-the-shelf 28.8 modems can be had for as low as $160 today, but if you don't want a maintenance nightmare, you'll buy rack-mounted modems, which cost $600 for the rack (holds 16) and $240 for each modem.

Connected to the modems you need a terminal concentrator such as a Livingston Portmaster or a Cisco 2511. We had more experience with the Cisco, so we bought one of those for $3,000; it can handle 16 modems at a time. Crunch these numbers and you get a total cost of $515 to install a new port. Then there is the monthly cost of $20 per phone line, plus $700 for the high-speed connection back to Cambridge. (That's the service that was $400 when we first moved here.)

One of the critical issues in making a profit is determining the number of ports you have to have per customer. The fewer you have to buy, obviously, the better the profit margin. Before the switch to Daylight Savings Time, we had one port for every 10 customers. But some of our customers were getting busy signals between 7 and 9 p.m., our peak period. As soon as the time changed, however, our busy signals went away. With the longer days, customers were now dialing in at varied times.

Today we have one port for every 14 customers and no busy signals. On the other hand, the average length of each call and the average number of connect hours used by each customer every month is steadily increasing, so we will soon need to buy more equipment to give our customers the same level of service. That money will come directly out of our bottom line.

Currently, we're making a few thousand dollars each month. But to show how that can quickly evaporate, I still recall how back in March, I bought $3,800 of some much-needed equipment using my personal credit card. I'm still paying it off.

Our solution has been to raise our rates, and to avoid the trap of flat-rate pricing. The way I figure it, flat-rate pricing encourages people to leave their computers connected to our modems all the time, which ends up blocking other users, forcing us to expand even faster. If I give people an incentive to waste my resources, I'm sure that they'll be happy to oblige.

Meanwhile, as this article shows, I'm still making my living by writing.

Copyright 1995, The San Jose Mercury News. Unauthorized reproduction prohibited.


The San Jose Mercury News archives are stored on a SAVE (tm) newspaper library system from MediaStream, Inc., a Knight-Ridder Inc. company.