If you can read this, you might want to skip straight to the content. Also, kindly take a moment to read my rant about Web design—especially if you’re wondering why this site looks a bit…dull.

Through the Contact Lens of Ecommerce

A dot-bomb in your eye

by Ben Goren


I’m not a big fan of disclaimers, but it’s only fair to include one in this case. The events described here took place some years ago, and I wouldn’t be surprised if I get some of the particulars of this story worng. Treat this as an illustrative example based on real events rather than the Gospel Truth. Regardless of the facts of events, names most certainly have been changed.

In the beginning…

Some years ago, after my time working in a tech support call center and well before the peak of the dot-com bubble, I was doing a fair amount of work for MacTemps, now Aquent. They treated me well and got me good work; I tried to return the favor as best I could.

One day, I got a call from Christine, one of the agents at the local MacTemps office, about a job: Richard, the owner of a small local graphic design firm, wanted to hire a Web designer for work on one of his client’s Web site. Before I followed through on the call, however, Christine had some warnings to offer me.

Richard had had some dealings with MacTemps before, and they didn’t speak well of him. He had a habit of wanting to pay bottom-dollar for the best talent MacTemps could offer. More than once he asked a temp to work for him directly, for less money, and without telling MacTemps?all of which is in violation of various contracts. They suspect he might have succeeded with one person. He had been late in paying MacTemps on one occasion.

Christine told me all this, and let me know that they were giving Richard one last chance. They wanted me to be on the alert for any funny business. If I got any, they’d back me to the hilt and end all future dealings with him. And Christine especially wanted me to let her know about everything that happened.

A fellow brass hole!

Forewarned and forarmed, I called Richard. We talked a bit about the project. One of his traditional print customers manufactures blanks for contact lenses. Opthalmologists order the blanks in bulk and grind them as necessary to match a patient’s prescription.

This company—call them ACME—already had a Web site on which they advertised their products. They had just introduced a new product, similar to one of their existing ones, and they needed a new page for that product (based on the existing pages) and some minor edits to the rest of the pages. Not a problem, I said. Richard also wanted to talk about some other, bigger plans when I got there; sure, I said.

Somewhere in that conversation, it came out that we’re both trumpeters. He’s more of an amateur, but still enthusiastic. He told me to bring my horns over and we’d play a bit. Of course, I said.

At last we meet

So, I went over to Richard’s offices to meet him and discuss the work. We talked trumpet mouthpieces and blew a few notes. The updates to the existing site really were no big deal; the biggest problem turned out to be getting the passwords to make the changes take effect. His bigger plans were…well…big.

Richard wanted to create a complete ecommerce site for ACME, everything run in-house at ACME’s facilities.

That’s easy enough to say, and a neat idea. Richard, however, clearly had no idea whatsoever was involved with such a project?at least, not the way that he wanted things done.

A quick tangent

It’s worth taking a moment at this point to note that, today at least, it’s not that big a deal to create a simple ecommerce site, so long as you don’t insist on doing everything yourself. There are numerous companies—Yahoo! springs to mind—which offer “storefronts in a box” where you interactively answer a few questions, supply some basic information such as product descriptions and prices, and, along with an appropriate exchange of money, you get a quite acceptable site.

Or, if you really want to have everything done in-house, there are many companies, from individual contractors like me to Big Blue itself, who will be happy to do everything for you…but you can expect to pay a pretty penny. The extra expense generally doesn’t make economic sense unless you’re doing an awful lot of selling or if you’re such a big company that sums this large can come out of petty cash.

The details


As Richard envisioned this, ACME would get a computer (or computers) to act as the Web server and a high speed Internet line. At the time, the computer would have cost a couple thousand dollars (at a minimum)…and the ’Net connection would have cost about that much every month. As anybody who has ever purchased computer equipment knows, there are frequently additional hardware expenditures needed that nobody anticipates until after the first round of equipment has been purchased. In this case, I later found that installation of the ’Net connection would have cost $10,000 at bargain-basement prices.

Once you’ve purchased a computer, that’s just the beginning. You need to set it up, install the proper software on it. After initial setup, servers need ongoing maintenance. All operating systems have had security vulnerabilities discovered from time to time?some more than others?and computers that store customer and financial information must be regularly updated to protect against potential attacks. Few computers are fast enough for everybody’s desires; careful configuration can devote a computer’s limited resources to those tasks where it’ll do the most good. Determining what need changing requires monitoring how the computer performs over time and under stress in the environment in which it functions.

Or, in other words, you need a systems administrator for your server. Systems administrators don’t come cheap; salaries of $80,000 are not unreasonable. Considering that this would be a single computer rather than several, ACME might be able to get somebody part-time or expect the sysadmin to wear more than one hat.


The server would host their existing Web site and take orders for their products. Hosting the existing site is no trouble; most any kind of server operating system (I was using Linux at the time, but Microsoft and Apple both had reasonable server options—and that’s ignoring everything from Sun, HP, and the other heavy-hitters had) could do that right out of the box. But processing orders would require a CGI program to deal withe information it recieves after a customer clicks the “Submit” button on the order form. CGI programming is exactly that—programming—as opposed to simple Web page design. It’s not uncommon for programmers to get paid as much as lawyers, although CGI programming generally doesn’t garner top dollar.

But what would the computer actually do with the information once it had been gathered? A simplistic approach would be to gather it all up into an email note and send it to ACME’s secretary, but that’s hardly using the equipment to its full potential. Why go to this trouble for something that, in the end, takes as much human involvement for customers and for ACME as a FAXed order? Far better would be to have the computer handle as much of the grunt work in daily business as practical.

For example, the computer should be able to verify payment without a lot of trouble. For credit card orders, the computer could be interfaced with the electronic transaction systems offered by the credit card companies; doing so might have meant some custom programming to make the connection work properly, but it certainly would have been practical. For customers with an existing account at ACME, the Web server could query ACME’s customer database for the relevant information.


But wait. Did ACME have a computer-based customer information and accounting system? If so, what was it?

The chances that they had an existing system that was easily networkable were remote. If we were lucky, they were doing everything on paper, in which case we could create the systems from scratch the right way. If we were unlucky, they were using some old DOS-based system, poorly designed to begin with, which only allowed one person sitting in front of the computer to do anything. Then, we’d have to set them up with a new system and deal with the mess of transfering their old data to the new system. Only if we were very lucky would we be able to use their existing systems without having to buy and install a new one. The new system would be comparable to the Web server. For security concerns it should be a separate system.

And what about inventory? It would be quite embarrasing for one of ACME’s customers to place and be billed for an order only to find out that they were out of stock. Did ACME have a computer-based inventory system? Did it suffer from the same problems as might have the accounting system? Fortunately, using the same computer for inventory and accounting would not be unreasonable, but that would be the only significant shared cost between the two. Programming and personnel expenses would again dwarf the hardware costs.

In short, ecommerce (in this context) only makes sense if your business is all run by well-designed computer systems. If that kind of infrastructure is already in place, you probably already have the resources to do ecommerce with a relatively minimal additional amount of effort. If you don’t however, you’re first looking at al the expenses of moving your business into the computer age. Are there great benefits to doing so? Certainly. Is it expensive to do so? Most assuredly. Is the cost worth it? Ask your accountant and financial analyst.

The wrinkle

As we were discussing all these sorts of details—and Richard didn’t have answers to any of them—we came across another bump in the road, so to speak.

Contact lens blanks are controlled medical devices, just like syringes, x-ray machines, and prescription drugs. The FDA won’t permit anybody except a licensed physician to buy them; even the doctor’s secretary isn’t allowed to authorize the order.

Oooooh boy.

Web pages start out by being relatively anonymous. The server can get some basic information about the remote computer requesting a Web page, but it’s not too useful and can easily be forged?and it has no information about the person sitting in front of that computer. We needed some way to be sure that the person sitting in front of the computer is a physician authorized to make the purchase.

Put simply, there’s no way we could have done that. We could have mailed passwords to all of ACME’s existing customers, but that wouldn’t have stopped somebody in the mail room from intercepting a password and using it maliciously. Besides, what’s the point of going to all this hassle just so their existing customers could have one more option for placing their orders? It certainly wouldn’t generate any new business.

We could have required that, the first time a physician placed an order, verification (probably in the form of a copy of the physician’s license) had to be sent through traditional means. But if you’re going to get the authentication through traditional means, why use the Web site for the order?


That last last is an important point. Having an ecommerce Web site is neat, but it only makes sense if it’s going to bring in enough (additional) income to pay for it. If you’re using one of the simple pre-cooked $19.95/month sites, that’s probably not hard to do—but you probably also can’t do more than run a small living-room business with one, either. If the site is going to cost you, say, a quarter of a million dollars per year (including salaries, benefits, etc.) to run, it had better be bringing in seven figures annually that you wouldn’t have gotten without it.

And the mere fact that you have an ecommerc Web site doesn’t mean that you’re going to magically start selling things on it. People first have to know that you exist, and then they have to be convinced that they should spend their money there.

In other words, you need to market your site.

For some reason, people at the time thought that the only way to market Web sites was through banner ads and search engines. In reality, nothing could be further from the truth, both then and now.

I’m not a marketer, but I’d guess that the only sensible way to drive traffic to this never-created ACME site would have been by hiring a marketing agency to place ads in trade magazines or send direct mailings. Or maybe they would have done best with a traditional sales force, or…like I said, I’m not a marketer, but I do know that opthamologists are going to be spending more time reading trade magazines than surfing random Web sites looking at banner ads.

And all of that begs the question, what do you get with the ecommerce site that you don’t with an 800 phone number? Heck, the 800 number and salary for somebody’s teenager would be less than the Internet access alone and would have done the exact same job.

En fin

In the end, the requirement for authenticating the physicians would almost certainly have been a show-stopper, but things never got far enough for us to worry about it.

A few days after I finished the simple Web page updates, I got a call from Richard; he wanted me to put together a proposal for ACME’s site, including my fees. As soon as we hung up, I called Christine at MacTemps; this was news to her, as she was under the impression she had processed the end of the paperwork for Richard and he had said nothing about additional work. Christine and I had no doubt that Richard was trying to hire me directly and undercut MacTemps, but she called him directly, anyway.

I don’t remember exactly the ending, who said what, but Richard was unwilling to pay me to put together the proposal. The scope of the project was such that the proposal alone would have meant a week’s worth of work—and he could easily have taken the proposal and handed it to somebody else as a job requirement. I wasn’t willing to do that kind of work for free, especially as it should have been billed out at substantially more than what he was paying for me as a Web page designer.

I went on to do more work for MacTemps, and neither I nor they heard anything again from either Richard or ACME.