Predicting the future of computing by understanding the past
Now that Thanksgiving is over I am ready to prepare to come up with predictions for 2010. This year, I decided to start by looking backwards. I first entered the computer industry in the late 1970s when mainframes roamed the earth and timesharing was king. Clearly, a lot has changed. But what I was thinking about was the assumptions that people had about the future of computing at that time and over the next several decades. So, I thought it would be instructive to mention a few interesting assumptions that I heard over the years. So, in preparation for my predictions in a couple of week, here are a few noteworthy predictions from past eras:
1. Late 1970s – The mainframe will always be the prevalent computing platform. The minicomputer is a toy.
2. Early 1980s – The PC will never be successful. It is for hobbyists. Who would ever want a personal computer in their home? And if they got one, what would they ever do with it — keep track of recipes?
3. Mid-1980s – The minicomputer will prevail. The personal computer and the networked based servers are just toys.
4. Mid-1980s – The leaders of the computer industry — IBM, Digital Equipment Corporation, and Wang Laboratories will prevail.
5. Early 1990s – The Internet has no real future as a computing platform. It is unreliable and too hard to use. How could it possibly scale enough to support millions of customers?
6. Early 1990s – Electronic Commerce is a pipe dream. It is too complicated to work in the real world.
7. Mid-1990s – If you give away software to gain “eyeballs” (the popular term in the era) and market share you will fail.
I could mention hundreds of other assumptions that I have come across that defied the conventional wisdom of the day. The reality is that these type of proof points are not without nuance. For example, the mainframe remains an important platform today because of its ability to process high volume transactions and for its reliability and predictability. However, it is no longer the primary platform for computing. The minicomputer still exists but has morphed into more flexible server-based appliances. The PC would never have gotten off the ground without the pioneering work of done by Dan Bricklin and Bob Frankston who created the first PC-based spreadsheet. Also, if the mainframe and minicomputers had adopted a flexible computing model, corporations would never have brought millions of unmanageable PCs into their departments. Of the three computing giants of the late 80s, only IBM is still standing. Digital Equipment was swallowed by HP and Wang was bought by Getronics. The lesson? Leaders come and go. Only the humble or paranoid survive. Who could have predicted the emergence of Google or Amazon.com? In the early days of online commerce it was unclear if it would really work. How could a vendor possible construct a system that could transmit transactions between partners and customers across the globe? It took time and lots of failures before it became the norm.
My final observation is actually the most complicated. In the mid-1990s during the dotcom era I worked with many companies that thought they could give away their software for a few dollars, gain a huge installed base and make money by monetizing those customers. I admit that I was skeptical. I would tell these companies, how can you make money and sustain your company? If you sell a few million copies of your software revenue will still be under $20 million — before expenses which would be huge. The reality is that none of these companies are around today. They simply couldn’t survive because there was no viable revenue model for the future. Fast forward almost 20 years. Google was built on top of the failures of these pioneers who understood that you could use an installed base to build something significant.
So, as I start to plan to predict 2010 I will try to keep in mind the assumptions, conventional wisdom, successes and failures of earlier times.