Printer Friendly Page E-Commerce, the Path Not Taken – Xerox, Citibank, Al Gore, and me

6.2  E-Commerce, the Path Not Taken –
 Xerox, Citibank, Al Gore, and me


This article is a historical review, written by Alan F. Kay with input from Charles Pyne, for the purpose of acquainting a select few persons with relevant background information to help them appreciate and understand the significance, credibility and potential importance of the patent: MARKET PLACE SYSTEMS PATENT APPLICATION by Alan F. Kay, and Charles Pyne.
I am intrigued by the cumulative significance of four little episodes from 1980 to1992 that at the time were simple disappointments. In each case on reflection immediately after the episode I believed that I or a colleague could have said something differently that would have produced a much more positive outcome. With twenty-year hindsight, I now believe that I might easily have done something in each case that not only would have improved the lives of many people but, more important, would have changed the course of the world significantly for the better -- a bold claim that this article will endeavor to make credible. The opportunities lost in all four episodes appear now to be a single opportunity that was evolving over time. Though the possibility seems slim, it may even be true that a great benefit will yet be recovered. First, a little background.

I was the founder in 1965 and until ’79 the CEO of AutEx, the first e-commerce company. AutEx operated computer networks that were on-line, real time marketplaces for various industries long before the Internet era.  As a computer applications expert dreaming up the AutEx concept in the mid ‘60s, I knew that three big developments would be required to make marketplace systems possible: first, being able to use communications lines leased from AT&T; second, acquiring a state-of-the-art large computer; and third, obtaining terminals with full screen displays, keyboards and a modicum of intelligence. When all three were available I knew we could design the systems and software that would allow us to develop an enormously successful e-commerce business. All three pieces fell into place about 1968 and our first system, block trading of equities by broker/dealers and institutions was revenue producing on August 1, 1969 and profitable shortly thereafter.

The first AutEx terminals were provided by CCI, Inc. in 1968. AutEx’s central computers included the satisfactory Sigma 5 and Sigma 9, purchased from Scientific Data Systems (SDS), a Xerox subsidiary. In that era Xerox was one of the most respected companies in America. It was not hard to persuade us to buy 100 custom terminals in 1971 from Scientific Data Systems (SDS), whose CEO, Max Palevsky had sold his company to Xerox in 1969 for over $920 million – up to that time the largest American corporate takeover.  Palevsky and three of his associates went on the Board of the parent Xerox and Max became the computer acquisition guru for the whole company. This was weird considering Palevsky’s own experience was servicing the scientific market, not the much larger business market that Xerox was aiming for.  The 100 that AutEx bought were expected to be the first batch of many more SDS terminals. The reality proved quite different. No terminals AutEx acquired were as unsatisfactory as the SDS terminals. They required too much desktop space and too much maintenance. They were a disaster for AutEx. That episode preceded by many years the greater disaster enveloping Xerox’s drive to become a major player in computers[1]. So lost opportunity #1 came from me not noticing this foreshadowing of the Xerox failure.

Ten years after its start-up, in late ‘76, I sold AutEx to Itel, at the time a rapidly growing conglomerate, and stayed on to run my “baby” as a division of the new owner. By early ‘79 Itel’s remarkable growth began a rapid collapse and its cash reserves were running low.   My interests in life had already started to shift, and I took the opportunity to resign. I had started two successful public companies and had enough money to buy my family a lifetime supply of anything sold on supermarket shelves. (I am ruling out yachts and Ferraris) I also noticed that, on those grocery shelves, you could easily choose from among fifty brands of soap or 100 different breakfast cereals, but you could not find good public hospitals, rail systems, or public schools. Clearly public policy could use help that would likely be more beneficial than my starting yet another company dedicated to creating individual wealth, not a wide range of public goods and services.

Since AutEx was one of the few Itel divisions still growing and profitable, I recommended that Itel sell AutEx to realize cash. Itel followed the recommendation, auctioning off AutEx to over 20 bidders. The successful bid by Xerox, was double what AutEx had cost Itel 2 ½ years earlier. My separation agreement with Itel required me to assist in the transition by consulting with Bill Senter, the Chairmen of the Xerox Publishing Group, to which AutEx reported. I had no desire to play a further role in owning, managing or operating e-commerce, but still was excited about AutEx taking on new challenges within Xerox. I was proud of AutEx and wanted it to go on to greatness regardless of who owned it. In his Stamford CT headquarters in ’79-’80, Bill Senter and I had good conversations that covered some possibilities for expanding and strengthening his new acquisition.

No one knew in 1980 that over the next decade the personal computer was about to stage the most rapid growth and cost/performance improvement of any technology ever. Neither I nor anyone else imagined that by the end of the decade, the enormous flexibility, power and cost reduction of the pc was going to be available for e-commerce terminals. With the limited intelligence of AutEx terminals[2], it was not possible to provide the service that many e-commerce customers wanted. One example had been burned into my mind. AutEx’s Corporate Bond Trading System, operated successfully in its first year in the early ‘70s. No central computer could compute bond yield calculations fast enough for the few hundred customers we had by the end of the first year, not to mention the thousands we expected. Distributed intelligence was required.[3] Other factors played a role too, but that limitation led quickly to the System’s demise with potential lost revenue estimated in the tens of millions.

I was not aware that Xerox PARC (Palo Alto Research Center) was proving out virtually all of the capabilities that would be needed to make a personal computer that by the end of the ‘80s would have as much capability and capacity as AutEx’s own initial central computer, the SDS Sigma 5. When I met with Senter we talked about the future of information technology, e-commerce and in particular what AutEx needed and could accomplish. At one point I happened to mention that, at the request of an investment banker friend who was considering a financing deal for Ray Kurzweil, I visited and reported on Kurzweil’s new company that produced voice-based computers for the blind and computers that simulated musical instruments out of keyboard controlled tone generators. When, without mentioning that he had any financial interest in Kurzweil, Senter asked me what I thought of Kurzweil’s operation, I gave an honest favorable answer. Later I learned that the Xerox Publishing Group had landed the Kurzweil deal. My investment banker friend’s bid was not competitive.

Now here is the missing piece that raises the question, “Whose side is Senter on?”. If Senter had suggested that I visit the folks at Xerox PARC, with an explanation like, “They are doing things that might help marketplace systems work better,” I would have gladly made the trip. Enhancing marketplace systems was close to my heart. I am sure I would have seen the significance of PARC’s goals and become further involved consulting with Xerox, as I did two years later with Citibank.

Why did Senter ask me about Kurzweil’s operation, irrelevant to AutEx and not speak about PARC, which was very relevant? The answer is unknown. It could have been this. Kurzweil and AutEx were both acquisitions of Senter’s own Group. PARC was controlled elsewhere in Xerox. My visit to PARC might have led to the opening of a big new market for Senter’s AutEx division. But that was remote in time and one step away from his immediate needs. Lost opportunity #2.

Next, Citibank. Around 1983 I ran into Alan MacDonald, a member of the Executive Committee of Citibank, recently appointed as its Director of Strategic Planning. He was very interested in my background in marketplace systems and asked many probing questions. He wanted me to be involved in Citi’s Information Division, then in formation. He brought me in to meet a few others on Citi’s Executive Committee for further discussions. When I suggested that Citi and a group that I would put together start a jointly owned company to develop marketplace systems, MacDonald reacted with “good idea”, and added, “You know how large companies have to follow their procedures.” He thought it best if I started as a consultant to Citi, bringing in others from my group as consultants. I accepted. I formed a new company, OnLine Markets (OLM) Inc. and brought in three other ex-AutEx people. Our job at Citi was to examine acquisitions or internally funded start-ups that would get Citi into the business of running marketplace systems for the financial industry capable of producing $100 million annual revenue for the Information Division.  After a few months of examining and proposing many prospects, it was clear to me that Citi would only go into that business with an established company that brought a lot to the table besides know-how. The OLM/Citi joint venture was not going to fly. In the three years that OLM spent as consultants well-paid by Citi, OLM’s design experience was used by Citi to start joint-ventures of Citi and a few other major (multi-billion dollar) banks and brokers to service new trading systems that were relevant to Citi’s existing experience and interests. No small or medium sized companies needed to apply.

The only acquisition Citi made in the Information Division was Quotron, at the time one of the big three stock quotation vendors, i.e. contract service providers. Quotron was unique as the only technology company that still designed and manufactured its own computers from scratch, not for sale. I was unenthusiastic about the Quotron merger, and so was not invited to the Citi meetings leading up to Citi’s takeover bid.  My attitude had nothing to do with Quotron’s uniqueness, but came from my AutEx experience of what it took to be a successful information provider to the securities industry. I believed that Citi would not know how to run such a business. When I saw who was put in charge of Quotron after the Citi acquisition, I had little doubt that the management would fail operationally. It did.

Out of the hundred or more business plans OLM put together for Citi, I recommended once or twice a smallish, but technically strong company, with an exciting new concept, that I believed could meet Citi’s rapid growth requirements. But Citi turned down any acquisition that was not a big name or joint venture with big names, a requirement never mentioned in all the discussions we had on what we were expected to do for Citi. Lost opportunity #3.

How could I have done better?  Simple, perhaps this. When Citi first made me the offer to work as a consultant, I could have played harder-to-get. I thought that such a hard line was not likely to lead Citi back to making a joint-venture deal with OLM, but perhaps not. We will never know.

The final story first appeared in my book[4]. In ’91-’92 my non-profit, the Americans Talk Issues (ATI) Foundation, was working with Rob McCord, head of a 501C-3, the Congressional Institute for the Future, to run a Public-Interest Polling (PIP) project to perform pro bono survey research for Congress with the purpose of finding out what people really wanted for policy and legislation on major national issues.

We circulated to all 535 Members a “Dear Colleague” letter signed by Al Gore and other members of the House and Senate (an equal number from both parties) to explain what the project would find that Members did not already know about where the public stood on issues and how Members could participate in the project and benefit from it. The staffs of Gore and a few other Members submitted questions that the public answered in PIP surveys. These questions were carefully designed to get valuable information on how far Americans would go in helping other countries in various circumstances and to contribute to solving the truly global environmental problems. The responses had come back with a clear and consistent viewpoint, and I knew Gore’s staff appreciated receiving the findings.

I first met Al Gore, a House Member from Tennessee, when he came to Massachusetts for the AAAS conference on the MX missile in the early 80's. The MX debate was heating up between those who felt the MX was a waste of money and destabilizing (thus making nuclear war more likely) and those who saw that it filled a niche role in the strategy that our deterrence posture required. The meeting was primarily the place for technical papers on one side of the debate or the other. By the questions Gore asked, it seemed clear that above all he was looking for a middle-ground compromise platform. He seemed to find it in the position that we should not build zero as the doves wanted or 200 as the hawks wanted but 50 MX's. While I understand the politician's need for compromise, in this case it bothered me that someone would come to a meeting of that depth and not examine the two sides a little more before opting for compromise.  I felt a good case was made for building none. I could understand that if you thought we needed this missile for some strategic purpose, then you would size the need and be in favor of building that number, but building a compromise number was like being a little pregnant. Gore had his position. He just wanted a rationale for holding the position. I wanted to shrug him off as a lightweight. On the other hand, he was one of the few, probably the only, Member of Congress to attend that conference. He was seeking the compromise point, not on the merits of the case, but on the politics. He was doing the best that he could under the constraints of The System.

In the Senate, he was Mr. Environment, quite literally. He was by far the best presenter in the Congress to a general or political audience on the cause and nature of some of the complex environmental effects of human activity, like the serious consequences and causes of the hole in the ozone layer and a potential major shift of the course of the Gulf Stream. As a candidate and later as Vice President he became almost invisible publicly in environmental matters, presumably because he was asked to do so as a "good team player." He was also a proponent of government support for the information superhighway. No, he did not invent the Internet, but he protected and assured it a transition from public to private ownership at no charge to the favored private organizations. In 1991 I tried to meet with Gore to discuss how the shift of survey research to cyberspace was likely at some point and the government should consider guiding and encouraging that as a tool of democracy.

Roy Neal, Gore’s Chief of Staff, asked for and got a one page summary of the purpose of the meeting and assured Rob McCord that he would work to get a meeting arranged. Instead the noted author and lecturer on global issues, my spouse Hazel Henderson, and I kept meeting Gore at events such as the UN Conference on Environment and Development at Rio in 1992, where Gore told me to consult his appointments secretary to arrange for a meeting with him back in Washington. No such meeting could ever be arranged. Rob shifted over to trying to get meetings with Gore’s environmental aide, Kathleen McGinty, a self-described "fan" of Hazel's. She canceled lunch three times. This was all occurring during the Presidential campaign. They were clearly busy, but there was more that came out a year or two later.

McGinty had become the environmental affairs aide to the new Vice President Gore. I was standing next to the doorway of a luncheon event honoring the new corporate endorsers of the CERES Environmental Principles.  As McGinty came from behind me, and entered the room she turned her head toward me for a moment. Just as I was beginning to recognize her, she said without really stopping, "Oh, Alan, we should have listened to you."  Before I could deal with a response, she slipped by into the room. That was all, I was in the moment. I put the remark aside. Three or four different possibilities for what she meant, mostly trivial, occurred to me later.  If she had something real to say, not chit chat, she knew my coordinates and could have communicated. She hadn't. After all the missed appointments and unreturned phone calls, there was nothing more I could do to communicate with her or Roy or Al. I had been through all that.

From May through July 1992 I attended a series of dinner Roundtable Meetings in Congress sponsored by a progressive organization, the Environmental and Energy Study Institute (EEI).  The meetings, organized by Gareth Porter and facilitated by David Gergen, attracted several Senators and Members, and a variety of some 25 foundation and policy organization people. The second meeting focused on the just completed Rio Earth Summit. The four or five in the room who had attended the Summit were asked to give their impressions. About to become Vice President, Al Gore, spoke first. I spoke third.

I spoke for five or ten minutes and my theme was that, as Rio revealed, the people, the NGO's, those from around the world and those from the US, were way ahead of the leaders on understanding and demanding action to face up to all the serious environmental issues. Supporting the argument in a natural way were the results of ATI Congressional surveys #19 and #20, which I summarized. I was elated because for the first time, some Senators finally had to hear how disconnected the Senate was on environmental, particularly energy, issues from what the American people wanted on a one-person-one-vote basis. I went home that night elated that finally Al Gore had to sit through and hear what he had seemed to avoid up to that point.

In some ways the penultimate Al Gore story was the most unlikely. I was attending a black tie annual affair of one of the big ten environmental organizations. Senator Gore was the featured speaker. A key purpose for my being there was to speak to several people concerning important matters bubbling up at GreenSeal, an environmental organization on whose board I served. The meeting was starting to break up. I thought I spotted one of the people I was looking for on the opposite side of the large, crowded room. I headed off, anxious not to miss her. As I was walking straight toward where I thought she was, I began to realize that I was also headed straight toward Gore who having come back down from the podium was standing in the middle of the room at his dinner table. He had just finished his speech, which was well received, but to my amazement, he was alone talking to no one. Usually after a speech like that there are a knot of people around the speaker. I cannot imagine why this was an exception. Gore seeing me walking toward him, clearly expected that I would speak to him and held eye contact as I approached. Here was a man whose ear I had diligently sought to obtain, largely unsuccessfully for over a year, and he was standing alone looking at me approaching, waiting to speak. Was God testing me or what? I made an instant decision. I wasn't up to trying any approach to the soon-to-be Vice President, but I had to say something. As I walked right by him, I tossed off, "Great speech, Senator," and walked on.

What should I have said?   I could have said that he should not approve the Internet as the major new communications system for the US and the world because it had serious design flaws. I knew of a better design that AutEx and others had used that could have served as the basis for a new network of networks much better than the Internet has done. The problem was that this was 1992 and I did not know how bad the Internet design was turning out to be. It was not until my colleague Charles Pyne and I had thought through our pending patent[5] that we understood what the fix could be. The fix will be expensive, time consuming, and disruptive, but if the Internet becomes a market failure for significant portions of the whole market, a better network of networks may turn out to be in demand fairly soon. If this is not the case, an improved network will evolve slowly with little disruption and annual investment costs minimized, spread over many years. What I could have said in 1992 to Gore was that the Internet should not be approved without considering its design flaws and how they can be remedied. If Roy Neal had not blocked such a meeting in 1991-1992 or if I had had the wit to think of that in the few seconds I had to introduce a new topic at the black-tie affair, instead of just passing off, “Great speech, Senator”, the world would be a different place today. Lost opportunity #4.

One thing all four opportunities had in common was the possibility of making personal computers and the Internet much more cost-effective. When it works well, information technology is magical; when not, an expensive nightmare. But the reason that the opportunities were lost in each case was a clash of motives and cultures (perhaps the “backgrounds” of the individuals involved is more accurate than “cultures”) capturing the mind-set of those involved, including myself, that led to the cutting-off of communication and discourse when it was most needed. The way the internet can be fixed, see 6.4 Marketplace System Patent Application.


[1] Douglas K. Smith, Robert C. Alexander, “Fumbling the Future: How Xerox Invented, Then Ignored, the First Personal Computer”, Morrow, 1988, and Michael Hiltzik, “Dealers in Lightning – XEROX PARC and the Dawn of the Computer Age”, HarperCollins, 1999

[2] Under keyboard control, the terminal had to be able to set a trigger to “yes” or “no” so that if the central computer sent a message asking if the terminal needed to send or receive specific information, the central computer could respond with whatever the system design provided.

[3] By the mid ‘70s AutEx had terminals with much more intelligence.  Driven by Data General’s Micro Nova computers fitting easily under a desk or in a remote closet, AutEx put monitors and keyboards on users’ desks in a combination that was fully programmable.  Compared to the pc’s available in the late ‘80s, the AutEx pc’s had four deficiencies: a) no disk drives, b) smaller memory, c) no color screen and  4) limited graphics.  None of these impaired the utility of the marketplace systems AutEx was then installing such as those needed by broker’s servicing government bonds, currency exchange markets, etc.

[4]  Alan F. Kay, “Locating Consensus for Democracy – a Ten-Year US Experiment,” 1998 Americans Talk Issues Foundation. St. Augustine FL. See #2 Public Interest Polling.

[5] “Marketplace Systems, #60/186,068, Alan F. Kay and Charles Pyne, 2/29/00.  See also “Marketplace System Fees Enhancing Market Share and Participation,” US 6,882,985 B1, 4/19/05

>>> 6.3 The Birth of E-Commerce

Table of Contents

Return to Top