Why Every Enterprise Needs
Web-Based E-Mail
by L.A. Heberlein, Seattle Lab
(Originally published in Messaging Magazine, May/June 1999)

Two and a half years ago, Web-based e-mail did not exist. Now almost everyone who uses a computer at all has at least tried Web-based e-mail, and for many new users, going to a Web-based e-mail account is the first thing they do upon accessing the Internet for the first time.

In Web-based e-mail, the user browses to a Web page, enters a user name and password, and a Web-based e-mail program running on the server displays the user’s mail on a dynamically-generated Web page. Nothing is downloaded to the user’s machine. No client software is required, so any user can use any system with Internet access to retrieve e-mail. This contrasts strongly with the previous model, client/ server e-mail. In client/server e-mail the user employed a client e-mail application (such as Eudora or Outlook). This client application logged in to a mail server—usually a POP (Post Office Protocol) server—using a stored user name and password and retrieved the user’s mail from the mail server, downloading it onto local storage on the client machine, where the user could read and manipulate it. Because of the configuration the client application required, and because the mail was stored locally, the user could easily read e-mail only at a single computer and would have difficulty managing e-mail from other locations.

The rapid spread of Web-based e-mail has been one of the swiftest changes in computer usage patterns we have ever witnessed. One oft-told story is the dramatic history of Web-based e-mail provider HotMail, which went from inception to a multi-million dollar purchase by software giant Microsoft in what seemed like minutes.

The first adopters of Web-based e-mail were public Web sites. Portals and would-be portals, Internet Service Providers (ISPs), community sites and affinity pages all sought to drive more traffic to their Web site and to give users a reason to return to the site. The primary economic support of public Web sites is banner advertising, and the goal of the sites is to increase the number of "eyeballs" that sees the banners. My company, Seattle Lab, was the first to sell a commercial Web-based e-mail product, and the first purchasers of that product were Internet Service Providers, expressing frustration that, as one of them put it to me, "people are using my dial-up accounts to log directly onto Yahoo, and never seeing my site at all."

The largest problem for most sites was not getting users to come to them for the first time, but providing adequate incentive to return. Web-based e-mail provided a reason for users to come back again and again. E-mail is something people use consistently. For many computer users, checking their e-mail is the very first thing they do every time they turn on their machine. So Web-based e-mail provided a powerful engine to support public Web sites. (The only equally powerful "return generator" is stock price quotes. Users will make frequent return visits to check stock prices. As widespread as stock market fever currently is, however, stockholders still make up a small segment of the population when compared to e-mail users.)

By 1999, Web-based e-mail has become an expected feature on public Web sites. Having Web-based e-mail on your site no longer provides a competitive advantage. Instead it is a disqualifying omission not to have Web-based e-mail. A major content provider told me just last week, "You can’t have a portal without e-mail. It’s a minimum requirement now."

But even from the beginning, we also saw Web-based e-mail installations at private Web sites. Web-based e-mail is strongly associated with free e-mail on public sites, but this association is mostly an accident of history, and does not reflect any essential connection between the two. The largest growth in Web-based e-mail at present and projected for the future is on private sites to which access is restricted.

The highest initial concentration of private Web-based e-mail sites is in education. When you consider the computing configuration of most educational institutions, it is easy to see why. At a school, there are more users than computers. Any user needs to be able to use any computer. Therefore, the traditional e-mail client is impractical. The e-mail client first expects to be configured with a great deal of information about who is using it—the name of the POP account, etc. Then, when the client connects to the server, it downloads the mail from the server to the client machine. In a configuration where the user may never use this specific machine again, this makes no sense.

Web-based e-mail, on the other hand, is ideal. The user can walk up to any computer, in a classroom, in the library, or in a computer center. With no configuration at all, the user simply launches Web browser, connects to a campus Web mail site, enters a user name and password, and checks mail. The mail is managed on the server. Nothing is left on the client. Another user may immediately use the same computer with no need to reconfigure it. The first user can use another computer next time, again, with no set-up required.

An alternative e-mail solution for educational institutions is IMAP. IMAP, like Web-based e-mail and unlike POP, can leave the mail on the server rather than downloading it to client computers. Not surprisingly, educational institutions have also been central vectors in the spread of IMAP, just as they have with Web-based e-mail. IMAP systems have their own strengths, which include a far richer set of functionality than most current Web-based e-mail systems. But this can also be a disadvantage, in the administrative overhead it imposes. This huge administrative burden and technological complexity is the reason that IMAP diffusion has been limited to the largest educational sites. A giant state university would implement IMAP; a smaller private liberal arts college, or community college, or high school, would instead turn to Web-based e-mail.

Web-based e-mail, in fact, can actually reduce e-mail administration, in some cases to close to nothing. At non-Web-based e-mail sites, administrators must create e-mail accounts before users can use them. Most Web-based e-mail systems can be configured to allow users to create and manage their own accounts. The system can also be configured to delete accounts that have not been accessed for a specified period of time. A typical community college, for example, faces huge quarterly turnover, seeing an almost completely new student body every ten weeks. Yet they can provide campus mail to this dynamic user base without requiring any e-mail administration whatever: new students create their own accounts; the accounts of former students just go away by themselves.

The benefits of Web-based e-mail are not lost on business people. E-mail has become an increasingly important part of everyone’s lives. For many of us, it is the primary way we stay in touch with our customers and co-workers. We grow acutely uncomfortable if it has been too long since we last checked our e-mail. Without getting our e-mail, we don’t know what’s going on. Our most important client might be waiting right now for a crucial reply. For workers who sit all day at their desks, this is not a problem. But among sales people on the road, or those of us out of the office for training or at conferences, Web-based e-mail is a lifeline. Even when I’ve been across the building working in someone else’s office for a couple of hours, Web-based e-mail gives me a way to find out what’s going on without having to return to my own office to read my mail on my own computer.

Because few of their individual institutions are yet providing them with Web-based e-mail, general users have turned in droves to outside providers. Most businesses are full of users who are using outside free e-mail systems. These users, particularly those who travel, such as sales people, use free Web-based e-mail accounts to stay in touch while they are on the road. I have not seen solid statistics on the use, but one IT director recently estimated to me that over half his users were making heavy use of outside Web-based e-mail.

While it might initially seem a good thing to shift one’s load off onto outside services—free services at that—IT managers have strong reasons to want to bring those users back into the fold. They worry about backup, archiving, monitoring, reporting, reliability, and security. Enterprise management is increasingly asking IT management to be able to locate historical e-mail, to analyze e-mail trends, and to answer such queries as "show me all e-mail to and from our largest customer," or "graph average response time to sales support e-mail queries." E-mail management tools are just emerging to allow IT managers to answer these questions. It is impossible to deliver meaningful numbers, however, when a large percentage of the company’s e-mail is actually transpiring through off-site third parties. Security is another potent concern. "There is probably nothing wrong," one IT director said to me at EMA’99 in Dallas, "in having my company’s most sensitive information sitting on some disk drive at some Web portal. But I would rather have it sitting on a disk drive in my server room."

Security, by the way, is another reason for adoption of Web-based e-mail at some enterprises, which have gone so far as to shut off POP access from the outside world entirely. POP—which in the majority of cases sends passwords over the Internet in clear text—is a known vulnerability, and is very difficult to secure. Web-based e-mail gives IT directors the choice of shutting off external POP access completely, and leaving http access (port 80) the only port open to the outside world. There are known-good methods for creating secure Web sites, which is why most of us now do not hesitate to type in our credit card numbers when ordering goods over the net. These methods can be applied to e-mail Web sites, giving mail a level of security that has never previously been possible.

When Web-based e-mail can offer users more flexibility—checking their e-mail from anywhere—without compromising security, in fact actually improving e-mail security, it is clear why Web-based e-mail is currently "crossing the chasm" from the relatively small pool of early adopters to the general enterprise. But there is more than this at work. Web-based e-mail deserves study not only for its own merits, but as an example of a larger paradigm shift, of which it is but the first instance.

Two rapidly changing conditions are driving the rapid adoption of Web-based e-mail: the ubiquity of Internet access, and the breakdown of PC niformity into a proliferation of device types.

We do not—most of us—yet live in a world where the Internet is always available, everywhere we go, all the time. But we can see that world forming. And, with it, we are developing an increasing expectation that we can get to the Internet from wherever we happen to be. Most of us do not yet have wireless modems in our laptops, or cell phones with screens and keyboards…but we have seen the prototypes, and we can tell what direction things are moving. Even a year ago, it could often be difficult to get on the Net during a 45-minute airport layover. Now there is almost always a way on.

This rapidly spreading omnipresence of the Net has contributed greatly to the rapidly growing variety in types of computing devices. For most of the last twenty years, almost 100% of computing has been at a single kind of device, generically called a "PC." And while models varied slightly from one to another, all PC’s were essentially the same. And yet, though they were all the same, because of their local storage, it was almost impossible for me to work from any PC but the one with which I was associated. My files, my schedule, my address book, the current state of all my work was on my PC. To a very real extent, it was my identity. (Which is why, for so many of us, the PC became a laptop, so we could lug it with us wherever we went.) To move to a new PC was a process that could require days, and leave one tentative about the ability to work for weeks afterward.

The PC is still with us, but it is no longer the only device around. Now there are Network Computers, Windows Terminals, Thin Clients, Handheld Computers, Palm Pilots, and increasingly computer-like cell phones and pagers. Televisions are getting disk drives, menus, and Internet access. New cars have more computing power on the dashboard than most businesses ran on when the IBM PC was introduced. Household appliances are getting CPU’s and IP addresses.

Looking at the swarm of new devices emerging every day, it is possible to envision a time not far off when the PC’s importance is only a memory. Rather than have our one "computer," we will be surrounded by a bevy of intelligent appliances. Each different in design, each specific in purpose. All with two characteristics in common:

• All devices are connected to the Internet all the time, and
• The user interface to all devices is HTML.

The first prediction is a truism that will seem like "duh" to those who come after us—why would you ever have any device that wasn’t on the Net? The second is perhaps slightly less self-evident, but, being in the software business, I can tell you that no one I know is working right now on any user interfaces not based on the Web browser. (Even Microsoft, which spent a decade developing its monopoly on the desktop user interface, a thing called Windows, is now running as fast as it can after the "Active Desktop," whereby you use a Web browser to operate your PC.)

Seen in this context, Web-based e-mail is not just a more effective way to support e-mail users today. It is the first example of a complete change in the way computing is done and managed. In the world of omnipresent connectivity and wide device variety, applications will move off the desktop, back to the server. For the same reasons mail administrators find it easier to manage e-mail on the server than to configure mail clients, IT staffs will prefer to manage application servers than individually configured applications on all those PC’s. Moving applications to the server will provide better management, control, support, reliability, backup, security, load balancing, and geographical independence. The economic analysis that forces change will cite lower total cost of ownership. But a side benefit will be removing burdens from the end-user. For much of the last two decades, users have lost focus on the job they were trying to get done while they dinked around trying to get their computer to do what they wanted. The migration of functionality from the desktop to the server will free the user from a job as system administrator of their own PC.

E-mail is the first application to make the migration from the client to the server, because e-mail is a driving need. It is the one that users demand the most. Other applications will follow—with "groupware" and "personal productivity" applications such as calendaring, scheduling, project management, and collaborative software first. And though we will be living in a transition state for a long time, it is possible to envision the death of client software, which has provided us with a phenomenal growth industry for most of the last twenty years. When all the functionality is on the server, and when the

Internet is everywhere all the time, the only client software you need is a Web browser—and every device will have one built in.  The shift from client/server to Web-based applications may not occur overnight. But I would not assume that it will take very long. Look once more at the spread of Web-based e-mail. Two and half years ago, Web-based e-mail did not exist, and now it is everywhere on the Web, is permeating schools, and has become an indispensable service most enterprises are offering themselves. I would not be surprised to see other applications migrate as quickly.    MM

Back to Table of Contents