Making Money on the Internet

by Philip Greenspun for the Web Tools Review

Note: this article has been superseded to a large extent by Chapter 2 of Database Backed Web Sites

Asking "How do I make money on the Internet?" is like asking "How do I make money on the telephone?" The Internet is a communications device; use it to communicate with people and maybe you'll make some money. Ignore it and you'll be about as prosperous as a company that doesn't answer its phone.

An example of a good commercial site is They took an internal database containing information that customers need (i.e., "the location of package #2389156") and made it available to everyone on the Web. You can nose around for their press releases if you like, but basically the entire site is functional. Does the site sell anything? Not really.

Examples of awful Web sites are too numerous to list, but prime candidates include

As of March 4, 1996, McDonald's was still waiting for this Internet thing to catch on and consequently didn't have a Web site. However, I'm sure that after they've spent their $millions, their site won't be half as good or 1/10th as trafficked as McSpotlight.

The Internet is a 26 year-old community. Businesses with billions of dollars that discover the Net and decide "let's see if we can make a quick buck off this Net thing" are probably not going to endear themselves to anyone. The people who will prosper are those that figure out how to provide a useful service that people in the community value.

Do something that is isn't done well in traditional media

Why does Travels with Samantha get 1/8th the traffic of Pathfinder and 1/3 the traffic of HotWired? Pathfinder is maintained by hundreds of full-time people; HotWired by nearly a hundred; Samantha took me two weeks to Webify. My theory is that Travels with Samantha, an illustrated travelogue, does something that is very poorly done in traditional media. Magazines print lots of nice pictures but don't want to print long text. Book publishers printing travelogues print long text but don't want to pay for color separations and printing for a lot of pictures. Pathfinder, by contrast, is mostly news. News is something that is done very well by TV, radio, news magazines (including Time-Warner's own), and newspapers.

Case Study: Canon versus Nikon

Canon and Nikon have been slugging it out in the high end camera market for years. The products themselves are fairly well matched, with neither system being dramatically better overall. SLR camera systems are too much of a niche market for huge TV campaigns so Nikon and Canon stick mostly to glitzy print ads in photography magazines. However, the consumer never really gets the information he needs from either company. They go to a camera store and someone who is completely clueless sells them two cheap zoom lenses that are good for nothing and they wonder "how come my kid with his $100 point and shoot gets better pictures than I do with my $1000 Nikon".

What if one of the manufacturers set up a truly interactive Web site? A consumer could send in a question like "I want to take pictures of trees but they seem to converge at the top of the photo" and the response would be "you need a perspective correction lens". The consumer probably didn't know that such a lens existed but might be very happy to pay $600 for one. Not enough people will ever want or need a perspective correction lens to warrant a full page ad even in a specialty magazine. The Web should allow Canon or Nikon to put exactly the right equipment in the hands of the consumer, resulting in much higher revenues for themselves and much better pictures for the consumer. Whichever company figures out how to use the Internet to help its customers take better pictures could probably pull substantially ahead of the other in market share.

[I wrote all of the above before Canon actually brought out its Web site. The result is far lamer than even I could have expected. They really deserve to go up there in the pillory with Kodak and Ford. Do they even have the owner's manuals for their cameras so that you can look up obscure features if you're planning a purchase or if you have a camera and can't figure out how to use it? No. In short, they have nothing that you couldn't find in their print materials and, in fact, less.]

What if my business is publishing?

If your business is publishing then you presumably can't just use the Internet to communicate with customers; you'll have to make money eventually somehow selling ads or subscriptions or information about your readers.

Communications and databases change a lot of business models. For example, most people would say that AT&T's best chance of making money is collecting profits from long distance service. Consider, though, that Joe Consumer uses an AT&T Universal MasterCard, an AT&T cell phone, and AT&T long distance service at home. Because of his cell phone calling pattern, AT&T knows that Joe drives by three travel agents on his way to work every morning. Because of his home long distance, AT&T knows that he dialed three airline companies, two SCUBA diving equipment retailers, and two Caribbean government tourist offices. Because of his MasterCard records, AT&T knows that Joe spent $7500 last year on travel. How much do you think those travel agents on Joe's route to work would pay for his name, address, and phone number? Probably a lot more than AT&T will make on Joe's long distance service.

The above scenario is merely suggestive and presumably privacy considerations will prevent some of the most obvious money making schemes, but information about how consumers are using information is going to be a publisher's biggest asset. This information can be sold to advertisers, used to deliver custom pages with articles and ads of special interest to each reader, and used to refine content to make a publisher more competitive.

A publisher that just puts up a bunch of static pages and doesn't try to figure out who is reading what is doing better than a publisher that isn't on the Web, but not by a whole lot in terms of getting business experience for the future.

The simplest way of figuring out who is reading your Web pages is to require registration. Aside from being inelegant, this probably cuts your readership in half because (1) users back out when they find they have to hassle with a form, username, and password, and (2) search engine robots won't be able to get into any of your pages so someone who uses InfoSeek or Webcrawler to search for "Costa Rica" won't find any of your pages on the subject.

A more subtle approach is to generate a session key ever time a new reader comes to the site and use the Netscape Magic Cookie protocol so that the reader's browser continues to tell you who is looking. Then come up with a contest or some other optional means whereby the user identifies himself and you can map all of the previous activity in the session back to that person. This requires some database and logging sophistication, but it is being done by some of the most forward-looking publishers (I know because I wrote their code).

sidenote for publishers

One of the most pernicious trends I've observed is the development of Web publishing ghettos inside publishers. This enables people who work in older media to forget about the Internet because "we have a division worrying about that." The fact of being on the Web is a detail. Having a special group for everything that goes on the Web is like having a special group for everything that gets printed on 4-color presses. A centralized group that helps everyone else in the company get their services up and running and possibly maintains some central relational database might be useful, but that group should be a service organization like MIS and not be responsible for all the content.

Don't expect to sell anything that the user doesn't already want

Selling physical objects over the Internet can actually work pretty well, if you have exactly what the user wants and an easy way for him to find it. Let's say the user wants a copy of The Marriage of Cadmus and Harmony to be shipped to a friend in Greece. Assuming their database engine isn't too slow, that means is about to make some money. Ditto for nerd hardware like disk drives and RAM.

However, I don't know anyone who has successfully attracted impulse buyers over the Web. People apparently find that strolling through a mall, bleak though that may be, is a less bleak and depressing activity than browsing commercial Web sites.

Despite its non-commercial nature, I can use my own site as an example. I get over 100,000 hits/day. The photograph at right is viewed by at least 500 people/day. If people will make a donation to a local charity (explained in the gift shop section of Travels with Samantha), I'll send them a print. To a first approximation, 0% of the people who see this image on the Web make a "purchase" decision.

I know what you're thinking: "Wake up and smell the roses, Greenspun, you're a bad photographer." I'd be perfectly prepared to accept that except that perhaps 5% of the people who see a nicely framed 20x24" print of this image choose to buy one on the spot (and the frame alone is $250).

A minor exception to this rule is that advertising and magazine art directors sometimes buy my photos after seeing them on the Web. Presumably their professional skill enables them to imagine how an image they see on screen would ultimately look in the pages of their publication.

My site actually does make a lot of money. I just don't get any of it. The manager of the Hotel Milvia in Costa Rica tells me that his nine-room hotel gets two or three bookings/week because of a recommendation in my travelogue. The businesses mentioned in all report thousands of inquiries from my readers, many of them leading to immediate sales.

Again, the lesson here is that you can sell people what they already want. People who read my Costa Rica story are about to go to Costa Rica. People who read my "where to buy a camera" page are very likely about to buy a camera.

Go forth and have fun

There's more to life than making money. That's not what the Internet is about. Try too hard and you'll just collect a few nickels plus high blood pressure. Do something fun, something useful, and maybe someday you'll get paid back in money or Karma.

[Note: If you absolutely desperately must make money on the Internet, I recommend starting a Web software company; I hear these Netscape guys are doing pretty well, for example.]

Return to Web Tools Review home

Reader's Comments

Like it or not, "banner advertising" seems to be the leading way to make money on the internet (perhaps only second to developing web sites). See for evaluations of dozens of different banner advertising networks, brokers, exchanges, and related resources.

-- Mark J. Welch, March 7, 1997
Add a comment

Related Links

Add a link