Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Have Another Cup? ]   |  Take My Job, Please  |   [ A Shot Across the Bow ] >>

Weekly Column

Take My Job, Please: How to Predict the Future and Become an Industry Pundit

Status: [CLOSED]
By Robert X. Cringely

Next week, this page will have a new look, I'm told. It will be sleek and modern, the frog will be gone, and all the high tech features that never worked quite right will be finally fixed. Or maybe it won't have a new look at all. That's the way it is with beginnings and endings; we know they are coming, but not always with precision. In the spirit of such imprecision, then, let's turn our thoughts to the future, and how to predict it without looking too silly.

Time eventually humiliates us all, but no one suffers at the hands of time more than prognosticators. Trying to guess the future is a game we nearly always lose. Where, for example, are our aluminum, two-ton, Buckminster Fuller Dymaxion homes, built in factories and delivered gently as snowflakes by dirigible direct to prepared foundations? Or forget the Dymaxion homes: Where are the dirigibles?

Experience suggests we'd be much wiser to predict not what the future will be like, but what it won't be like. Still, the so-called experts continue to make predictions. I've done so myself, hoping at the same time that people will forget what I've written — or at least that I was the one who wrote it — before reality closes in and makes me look too stupid. And as long as there continues to be a demand for predictions, which is to say as long as markets exist and magazines are published, experts still will be making them.

The lure, of course, is what the lady at the bank called the "miracle of compound interest." Whether it's a market shift or the next technical revolution, the earlier we anticipate some fundamental change, the more we will be able to benefit from that change. At least that's the theory, but there are downsides to being early, too.

So, if prognostication is unavoidable, then let's consider the rules that should underlie predictions. That way we can either have a better sense of when someone else's view of the future is likely to be wrong (which is to say, most of the time), or we can do a better job than most at making our own predictions. Here are the five rules I follow. Your mileage may vary.

1. We tend to overestimate change in the short term. While change is inevitable, and sometimes a mountain of change can be seen distinctly across the valley, few of us are good at measuring the width of that valley. We just naturally tend to see change as closer than it actually is. The best way to measure this effect is to sing along with a recording of Frank Sinatra or Ella Fitzgerald: We always hit the next note before they do. We can't help ourselves. Applied to the real world, this tendency has us predicting market peaks and selling out too soon, or predicting that some technical advance will change our lives long before it actually does. Bob Taylor, the Defense Advanced Projects Research Agency official who commissioned Arpanet, which ultimately became today's Internet, says that when he made that decision back in 1966, he expected the network to look pretty much like today's Internet by 1980.

2. We tend to underestimate change in the long term. When change finally does happen, it sometimes looks very different from what we expected. The men and women in white coats who ran mainframe computer centers in the 1970s and '80s weren't surprised by the invention of the personal computer, but they generally didn't see that PCs would change the market for mainframes. Back in those early Arpanet days, Taylor may have seen the development of most of today's Internet services, but he admits he never anticipated the World Wide Web. At this moment, full-service stockbrokerages like Merrill Lynch are reacting to the challenge of electronic brokers like E-Trade by giving their human brokers a web presence, not realizing that the game is already lost and most human brokers are doomed.

3. The more specific a prediction, the less likely it is to be correct. Predicting that stocks will go up is generally safer than predicting that a specific stock will rise (unless that stock is Microsoft). It's easier to say that consumers will buy higher-capacity removable data storage devices than to say they'll buy those devices from Iomega rather than from Imation, or from some outfit not yet heard from. But the "not yet heard from" category ought not to be rejected out of hand since that's where the real danger lies for current market leaders. Bill Gates worries far less about an Oracle or Sun — competitors he knows and believes he understands — than he worries about the startup that will inevitably clean Microsoft's clock. His nightmare has no name.

4. Past performance is a predictor of future results but not a good one. Statistical regression predicts quite reliably that whatever mutual fund manager had the best performance last year won't have the best performance this year. That's because there is a tendency to stick with what worked last year even when it stops working this year. He or she can do well, but it is very hard to stay at the top of any game. This applies to technology, too. Those who have had success with a particular technology will tend to stick with that technology until it is too late. This describes Prime and Wang in minicomputers. Today's hard-disk drives are likely in the long run to be supplanted by some new type of storage mechanism, and it is unlikely those devices will be built by a Seagate or a Quantum.

5. The most reliable predictions are those that follow established trends. George Soros's big killing in the currency markets several years ago was based on his understanding of human nature and nationalistic, rather than economic, interests. Knowing that the Bank of England would defend the pound sterling purely because that's what the bank saw as its duty, Soros bet that trend would continue and used it to quickly make more than a billion dollars. More recently, a consolidation is taking place among networking companies in which giant telecommunications outfits are buying up smaller data communications outfits. This trend suggests that Cisco, Lucent, and Northern Telecom are predators while all the other companies in that sector are prey. Following the trend means buying a basket of these latter companies. Find the trend, but don't sweat the details or the timing because you'll always be wrong.

Even with rules to follow, prognostication hardly ever works. The best we can hope for is to be right more than we are wrong. If we are right early enough to compound that advantage, the reward can be a Microsoft or a Cisco to balance our fleets of dirigibles and tracts of Dymaxion homes.

Comments from the Tribe

Status: [CLOSED] read all comments (0)