Sometimes the citation has to be more serious, though. When the right tone is businesslike, I put away the thoughts of Chairman Murphy and look to other sources. If the subject calls for a really stirring line, it's the generals and politicians to whom I turn.

John F. Kennedy got off a lot of great utterances in his abbreviated life. He once modestly described himself as "the man who accompanied Jackie Kennedy to Paris." But the line for which he will forever be remembered, the one that will perpetually guarantee him a $200 space in the "Quotations" category on Jeopardy! comes from his January 20, 1961 inaugural address: "Ask not what your country can do for you--ask what you can do for your country."

Even at a distance of 37 years, it's an electrifying call to action, evoking patriotism, pride and the nobility of sacrifice for the common good.

Which brings us to the subject of this month's column.

Chicken Little Only has to Be Right Once

I get the same wicked laugh out of the Bastard Operator from Hell stories as everyone else who has ever done any meaningful user support. Even so, I try never to forget that, irksome though walking them through their little crises may be, those users are the reason we have jobs. Even if we don't personally have any responsibility for their support, the income they bring to our businesses keeps us in business.

In my mind, that means it is incumbent on ISPs to cater to the needs of their users. You should devote as much attention as possible to how you can make your users' online lives as easy and pleasant as possible, because they're the reason your business exists.

Note that making users' lives easy and pleasant may very well mean that your job becomes more complex and demanding. That's why the title of this month's offering is "Ask not what your users can do for you..." Nobody outside your family cares how hard your job is--and, if you take your users for granted, there are dozens or hundreds of other ISPs who are eager to woo them away.

Many ISPs concentrate on competing on the basis of price, connection speed and the ubiquity of their POPs. Those are all important considerations, of course, but they're sure not the only ones. In some ways, they're not even the most important ones.

I get a fair number of inquiries from end users--many of them small business owners--asking me to recommend an ISP. The ones who are just getting their feet wet always want to know how much it will cost them to get set up on the Web. They may not (and usually don't) know exactly why they want to "get set up on the Web," but they know that's what they want to do.

They're desperately afraid of getting ripped off. Or left to figure it all out by themselves.

And they're not sure which is worse.

They don't understand the difference between hosting their own web server, collocating a server at their ISP and running a virtual domain on an iteration of their ISP's web server. They don't understand DNS or IP networking, they're lucky if they've even heard of ISDN, T-1s or Frame Relay and they assume they'll be able to get a high-speed connection installed on one or two days' notice.

Even worse, they have no appreciation of how much it will cost them to build a suitable web site--or even what that means, exactly. They want it to cost nothing and be done yesterday and they want it to be precisely what their business requires--all without having to put any real thought of their own into it.

Fear and loathing? Hunter S. Thompson has nothing on these guys.

And they've got every reason to be very afraid.

An Ounce of Image Is Worth a Pound of Performance

Local news radio stations in the San Francisco Bay area regularly run commercials for seminars on "the exciting new field of Internet consulting." The hucksters who promote these foul things assure their targets that "you don't even need to own a computer!"

P.T. Barnum, meet the brave new Internet.

The problem isn't just that the products of these seminars are the deaf leading the lame. The problem is that they're almost guaranteed to create a disaster for their "clients" and blame it, if at all possible, on you.

After all, it can't be their fault--they're "experts," right?

At the same time, your competitors are wracking their brains to come up with ways to entice your users away. They're giving away software, running contests and promotions and doing everything in their power to outsmart you and capture your market.

Your most useful counter strategy for both kinds of problem is to create a truly transparent set of end-user services and make sure that they're prominently featured both on the top page of your web site and in your advertising (you do advertise, don't you?) Don't just sell Internet access--create a package of services for small businesses and sell it. Do the same thing for businesses as you should already be doing for individual users.

Make it easy for them.

One popular strategy for making ISPs attractive to individuals is to bundle a collection of basic Internet access tools with a subscription to your service. Whip up a CD-ROM with Netscape Communicator or Internet Explorer for Windows and Macintosh, add in Eudora Lite for Mac and PC, throw in WS_FTP, WS_FTP32 and Fetch, give `em mIRC and Ircle and don't forget WinZip, UnZip and StuffIt Expander and you've added instant value at minimal cost and with almost no effort. The only drawback to this strategy is every other ISP on the planet is doing exactly the same thing.

If you want to stand out from the crowd, you're going to have to put out the additional effort and expense that will enable you to provide the extra value that will set you apart. That means providing ancillary services, such as roaming connectivity, chat, streaming media servers and so on.

It also means furnishing your users lots and lots of help in the form of tutorials and step-by-step instructions for installing and troubleshooting applications. It means thoughtfully designing your own web pages so that help is easy and obvious to reach from anywhere on your site. Above all, it means choosing to recruit and train a sufficient number of help desk personnel that, when all else fails, your users can get a human to help them solve their problems.

And it means keeping your help desk open evenings and weekends.

When All Else Fails, Follow Instructions

To capture the small business market, you'll need to create a similarly friendly and comprehensive set of services. In essence, you're going to have to create a package that includes an appropriate combination of services, training and consulting for various small business needs.

Start with education and tutorials. I firmly believe that the more you teach your prospective customers up front, the greater the confidence they'll have in what you propose. Create or commission a set of materials that explains the difference between customer host location, collocation and virtual hosting. Write up an explanation of different connection options and how long your customers can reasonably expect to have to wait to have each one installed. Tell them what each option will cost and make certain it's very, very clear how much of that price goes to the local telco and how much goes to you. Explain the difference between a passive web site, a web site with hooks to a back-end database and one capable of full-on electronic commerce not in terms of the technology involved, but in terms of the difference in cost, complexity and delivery time for each option.

Then develop a package for each of those options. Make alliances with artists who have enough experience to understand the effect that the complexity and bit-depth of graphics has on a web site's download performance. Have them set package rates for turning existing business logos into digital graphics and a la carte prices for other custom graphics. Cut low-cost deals with HTML hackers to build basic web sites from templates in a package like NetObjects Fusion or FrontPage.

Recruiting honest-to-Pete programmers to do the coding necessary to create back-end hooks to customer databases is a non-trivial challenge. It's a seller's market, and good programmers are a precious commodity. However, products such as Everyware's Tango Enterprise and Macromedia's Backstage, both of which run under Windows NT, reduce the challenge to something a considerably less-experienced programmer can handle. And practice not only makes perfect, but solving a problem once reduces the amount of wheel reinvention each additional version of the same problem requires.

Building a full-function Internet commerce site is an even bigger undertaking. AbleCommerce offers a package called AbleCommerce Builder for under $2,000 that reduces the task to a somewhat more manageable size. AbleCommerce Builder includes Allaire's Cold Fusion RAD system, Cold Fusion Studio, WebTrends log analyzer, ProtoNet's ProtoFax Internet fax and paging application and DCSi's Windows 95-based Font F/X 3-D text renderer all in one bundled package--although support for the constituent programs comes from their respective parent companies. There's also Everyware's under-$2,000 Tango Merchant system, but it only runs in the PowerMac environment at the moment, so most ISPs won't have any use for it.

Of course, you can always go to Netscape for purely Unix-based commerce tools, but, since Netscape's own developers believe in the Unix way, its products are based on the assumption that you're comfortable doing a whole lot of C and/or Java programming before you'll produce anything like a usable result. On the other extreme, you can head over to IBM's Lotus division and let them sell you their Domino Merchant package--a product so easy to use that they don't mention anywhere on the site that it runs only on OS/2 Warp (which you must purchase separately).

Hey, it's a division of IBM, remember?

If Nobody Uses It, There's a Reason

I also happen to think you should permit your business users to run CGIs and Java applets on their sites, even if they're doing it on virtual iterations of your web server. Yes, I know there are non-trivial security issues involved, but Rule #1 has always been "The customer's always right," and, damnit, customers want active content. Remember, you're trying to please them, not your system administrator.

If he/she complains that your customers are endangering your system by insisting on active content, point out Rule #2: "When the customer is wrong, refer to Rule #1."

I'd allow MIME file types such as Macromedia's Shockwave and RealAudio streams, since Shockwave always did allow and RealPublisher 5.0 finally permits all non-live Real media types to stream via vanilla HTTP. Sure, adding audio and animation negatively impacts the performance of your users' sites (and your system). But, once again, it's what they want.

I also think you should be as generous as possible with user disk space allotments. Let's face it. Disk space is nearly as cheap as dirt these days. There might once have been some excuse for being stingy with it, but those days have gone the way of $75-per-megabyte RAM.

And, for pity's sake, don't oversell your bandwidth to the point where your customers' web site performance is impaired. I fully understand that telcos overcharge for T-1 connectivity, but a reputation for poor quality service is the single most expensive possession you can earn. If your hose is saturated, even at peak times, upgrade that puppy.

Finally, keep after your users for input on services and policies they'd like you to provide. You can't afford to wait passively for information to come to you and they are absolutely your best source of data on the changing demands of the market.