Seeing eWorld show up on the list of Apple’s biggest flops was interesting to me. I understand the perception but having spent 4 years in Apple’s Online services gives me some insight into the rise and fall of eWorld that the author may not have possessed.

In 1992 Apple was trying to phase out it’s AppleLink system (run by GEIS, GE’s online service) with what would become eWorld. Apple looked at the major online services of the day to buy the technology to build the new system. They looked at Prodigy, CompuServe (At the time CompuServe still thought that online addresses that looked like 524364.243 were fine and that accounts like “chris2x” were in their words “vanity names”) and eventually settled on America Online. That in itself was ironic because AOL had originally been built as a joint venture between Quantum (later AOL) and Apple and had been called AppleLink Personal Edition. Apple had gotten cold feet about going into the online service industry and had backed out, but now with eWorld was returning to a partnership with AOL.

At the same time that Apple signed the deal with AOL (Fall 1992) there began a period of exponential growth in AOL’s membership. At that time that AOL had about 300,000 users. AOL started putting CDs in your magazines and cereal boxes and the membership grew rapidly. And AOL started to break. AOL’s engineer group scrambled to deal with the rapid growth in subscribers and server load. In part the money from Apple for eWorld helped fuel the marketing blitz, but this worked against Apple as it delayed engineering needed from AOL to meet the eWorld schedule.

At the time of eWorld’s development Apple Online Services was moved into a new division at Apple called the Personal Interactive Electronics Division (PIE) headed by Gaston Bastiaens. The more well known product that was also put in this division was the Newton. What became the Newton MessagePad was about to be announced. It supported infrared beaming and faxing but noticeable absent was email. Neither eWorld nor Newton were launched but the decision was made to loan a programmer from the eWorld team to the Newton team to add email. I was that programmer so I remember that part of the story well. Gaston used to walk into my cubicle in the Newton area late at night and ask in his Belgian accent “will ze email be in ROM?”. Will the email be finished in time that it can be built into the machine. I added email to the Newton beta by reverse engineering the AOL protocols with a network analyzer. AOL was not entirely happy with that approach and insisted that we move from one of their older technologies (FDO88) to a new one (FDO91). They also had two engineers on their side assigned to develop a new registration system that would be appropriate for the Newton. eWorld systems engineering resources also had to be redirected to NewtonMail instead of the eWorld launch to meet the NewtonMail schedule. Thus by doing my job of adding email to the Newton I helped slow down the eWorld launch. The irony was that when eWorld later folded there was no email for the Newton, when in fact NewtonMail was debugged against the AOL servers not the eWorld servers. Only one parameter change would cause it to work with AOL if AOL had been willing to turn on the appropriate server components.

The biggest effect of the eWorld delays was that there were at that time key dates that had to be met in order to ship eWorld bundled on every Mac. eWorld was not bundled on Macs for almost a year after it launched because it missed a key bundling milestone. Also because Apple’s manufacturing processes in those days did not change rapidly the eWorld client was included only Macs for at least a year after there was any eWorld service to which to connect.

When eWorld launched it cost about $8.95 a month with two free hours + $4.95 / hour ($7.95 during peak hours). The way that online services like AOL worked in those days is that they made money from people who had an account and never used it and from people who used it for many hours. People who used the system for 2 hours a month actually cost the company money because of the high cost of bandwidth. Those numbers seem outrageously expensive today, but were about the same price as other online services at the time and significantly less than AppleLink. The prices were dropped as time went along, although I don’t recall what those prices were.

The biggest factor that lead to eWorld’s demise was the rise of the world wide web. We must remember that back when online services were king in the pre and early web days, 2 hours of online time a month was fine for most people. After all, where would you go? Only 30% of people in online services tended to chat, so for many people they were just picking up their email. The web changed that. There were places to go. eWorld added a web browser in the eWorld 1.1 client version because it needed to do so in the worst possible way. And thanks to the browser provided by AOL, that was pretty much the way that browser worked – in the worst possible way. I was the manager of the Mac 1.1 client. We new the browser was bad, but had little choice. AOL was not fixing the bugs we were reporting and we had to have some client out with a browser sooner, not later. Since AOL was using the same browser in its software we certainly hoped they would eventually fix the bugs and within a few versions on the AOL client they did dump that initial browser (WebShark?) and build in Microsoft IE.

Another critique of eWorld was that there was no Windows client. Both a Windows client as well as a more full-featured Newton client were being developed. I know because I also managed those projects (how I got to be managing the Mac client, the Windows client and also working on NewtonMail at the same time is another story). The Windows client was about to go into Alpha when it was cancelled in August of 1995.

And here is what really killed eWorld. What we forget about those days at Apple was how dark they got. This is back before the return of Jobs and Apple is losing money. Some engineers in AOS’s Boulder Colorado lab had created a prototype of an eWorld that ran entirely on internet protocols. It used IRC for chat, NNTP for discussion boards and web pages for other content. Peter Friedman (the head of AOS) was able to get Apple’s executives to agree to keep eWorld alive but he agreed to scale down the operation significantly and move to an internet based service. But then the bottom dropped out of Apple, Inc. Apple was looking at over a $700 million dollar loss at the end of 1995. Whether you liked eWorld or not, huge cuts had to be made. eWorld was not profitable and would have to go. I have never faulted Apple’s executive staff for that decision. They had to save the company. They had to put their money in the Mac and let everything else go. So an internet version of eWorld never saw the light of day. eWorld closed its doors on March 31st at midnight.

There are two historic footnotes worth mentioning. One, a new company called LiveWorld opened an internet community called Talk City at the same time that eWorld closed with some of the same community leaders. I still work at LiveWorld for Peter Friedman, although we sold the Talk City site a few years back.

The other odd historic footnote about eWorld is that Apple made money on it. “Now wait a minute”, you say, “previously you mentioned that eWorld was not profitable”. That is correct. But when the deal was struck with AOL to build eWorld, Peter got AOL to include warrants to buy their stock. AOL stock before the deal was around $12 a share. You may recall that AOL stock eventually got to be so valuable that AOL used it to by Time Warner. When Apple eventually exercised its warrants it made more money than it spent developing eWorld.

Other eWorld links: