Monday, May 19, 2008

The Future

I've been thinking about "The Future" (capital T, capital F) lately, and wanted to consolidate my thoughts a bit:

1) There are no facts about the future. My friend Jeff Wacker, a futurist for EDS, pointed that out to me once and it really stuck in my head.

2) Long-term strategic planning is kind of silly, on several levels. First, there is very little accountability, because the people making the plans generally won't be around to see them come to fruition and won't be exposed to the consequences of being wrong. Second, there is precious little opportunity to learn from experience, particularly for long-range planning, because we don't directly experiences the results of our decisions and predictions - and if we do, it's too late.

3) We tend to overestimate what technology can do in 10 years... and underestimate what it can do in 50 years. I read that somewhere - don't recall the source.

4) When we talk about the future, we are really talking about the past. When we talk about the distant future, we're really talking about the very recent past. (That is, we are extrapolating trends, and the more recent the trend, the further into the future we think it will extend.)

4.5) Futurists don't really believe in change. That is, they think the future will be a certain way IF trends A, B & C continue and the environment/market/threat/etc responds the way it has in the past. These predictions basically rest on the assumption that things will continue, unchanged. This is not necessarily a bad assumption.

5) 20 years is a long time. It might make sense to spend 20 years on a research project, discovering some new knowledge, but a 20 year development project to build a new thing (particularly a large one) tends to deliver some components with obsolete technology while simultaneously relying on / hoping for some new technological breakthrough for other components, which may or may not happen (see #3 above). Never mind that the need/market/mission/threat has probably changed significantly over the 20 year development time (see #1 above). Perhaps that's why the DoD and the GAO both say we shouldn't spend more than 5 years developing something (the key word being "should").


Rhet said...

Having spent the last year in a futures studies / strategic planning program, I take exception to some of your assertions. I'll address them by number, as you've listed them.

1) True...there are no facts about the future. But there are educated guesses and fact-based assumptions, both of which are valuable.

2) Your assertion that "Long-term strategic planning is kind of silly" is particularly alarming, especially in the realm of defense acquisition. While true that we have no facts about the the world 20 years from now (or even 5 years from now), there is great value in trying to understand the range of possibilities that the future might hold (scenario planning for example). Without this "silly" exercise, we're left struggling to react. On the other hand, if we do some scenario analysis and long-range planning, we at least have an idea of what the future might hold and can plan accordingly. We won't get it right but we'll be closer to the mark than if we simply wait to react. As Douhet said, "Victory smiles upon those who anticipate the changes in the character of war, not upon those who wait to adapt themselves after the changes occur."

3) Based on my experience, we overestimate what technology can do in five years and underestimate what it can do in 10.

4) I don't believe this assertion would be true for most futurist (those trained to study the future). It is probably true for those who are unfamiliar with future studies.

4.5) The biggest problem with trend analysis is that people tend to extrapolate trends linearly and most change doesn't happen in a linear fashion. It goes through periods of accelerating change and then periods where change may slow or even halt (producing an 'S' curve). I believe that most technology today is on a rapidly accelerating part of the curve which makes forecasting extremely difficult if one doesn't understand the rate of change (and the change in the rate of change).

5) Yes, 20 years is a long time. But there are lots of systems that absolutely cannot be built in 5 years if they'll meet the needs of today's warfighters (I'm writing this with my defense acquisition hat on). Then again, waiting 20+ years for a new airplane is too long because so much does change. But constraining major weapon systems to 5 year acquisition cycles isn't realistic. Can we do better? Absolutely! But don't put artificial constraints on the system. We need better tech forecasting, not artificially constrained timelines. It's easy to poke fun at a weapon system that takes 20+ years to build--and to write cute stories about Krog building something he doesn't need--but that grossly oversimplifies the issue. Did the original reason for building the F-22 change over its development timeline? Sure. But will we need the capability of the F-22 in the future? ABSOLUTELY! Anyone who says otherwise doesn't understand the world in which we live. And we'll have it available when we need it because we have gone through the painful process of building it. We built an absolutely amazing system and because it was "over-designed" we'll be able to adapt it to all sorts of unknown future scenarios. That's exactly what we've done with all our existing systems. Ten years ago, no one imagined that we'd be using F-16's and B-52s to provide 'aerial artillery' for ground troops but that's what we've done and we've done it amazingly well. But we were able to adapt because we had a great technological toolbox upon which to draw.

The other issue you fail to discuss is money. A lot of the timelines in our acquisition system are necessarily driven by money. Could we build something in five years that does an adequate job for today's requirements? Sure. But could we afford to build an entirely new something else in five more years because the one we built this year doesn't meet the requirement and isn't adaptable? I'm not so certain about that.

I think the answer lies somewhere in the middle. We need to do some future studies and strategic planning but we also have to be able to quickly react and adapt. Importantly, this debate is happening at the highest levels in the DoD.

The Dan Ward said...

Well Rhet, I guess we just disagree on this stuff.

I agree future planning is necessary and can be useful, but I also think it's silly precisely because there are no facts and there is so little accountability and so little learning involved. Without facts, accountability and learning, what are we really doing? We're guessing. So I think it's important to recognize the inherent silliness in the exercise, and not take it too seriously. It is not the rigorous & scientific exercise many futurists make it out to be. Long term predictions aren't that different than fortune telling, in my opinion.

Now, educated guesses can be correct, but so can random guesses. For example, several experiments have shown that chimps can "predict" the stock market just as well (or better) than the experts. Yes, those results are disputed by the experts, who are also experts in explaining away their failures and shortcomings - but I haven't seen those results explained away by objective, independent reviewers. Probably because those independent reviewers are too busy laughing.

AS for #4 & 4.5, the only thing a futurist has to go on is the past - and the belief that past trends won't change. Even an s-curve projection is based on a belief that past s-curves predict future s-curves. Sometimes they do - but only if the s-curve pattern doesn't change. Thus my assertion that futurists don't believe in change (linear or otherwise).

I'll also have to gently disagree with your assertion about how easy it is to write a "cute story" like Krog. That wasn't easy, neither was it easy to get it published. Feel free to give it a try.

As for the F-22, well, that's a whole other topic... For now, I'll just say the US needs to accomplish certain effects. The F-22 is one way to accomplish those effects (a very expensive, long-delayed, overly complicated way). I think we could have established the necessary capabilities much faster and for much less money... but like I said, that's a whole other discussion.

Gabe said...

Some of my answers:

5. The point that the F-16 and B-52 are examples of systems that have morphed to serve other purposes is true, but only because they were developed and fielded quickly enough to determine what they could actually do. Remember, the F-16 was developed and fielded well under schedule; less than 10 years I believe. This point supports Dan's thesis.

I disagree with this statement:

"But there are lots of systems that absolutely cannot be built in 5 years if they'll meet the needs of today's warfighters"

If a system currently being developed cannot meet the needs of today's warfighter today then it's useless. Predicting were technology will be is tricky, but it's far better to exploit what we have today than to pin our hopes on tomorrow's breakthroughs.

2. Ricardo Semler, the famous CEO of the company SEMCO, thinks long range planning is silly. He "aids" his company in looking out 6 months and 5 years, but he doesn't put much stock in the 5 year look because he knows at the next 6 month interval, the direction will most likely, completely change. He essentially believes long range planning is not that useful(and 5 years is long range planning to him), because it often ties people to a particualr vision and prevents them from having a mind of adjustment. His explanation for using the 5-year interval is simply as an exercise to explore "the range of possibilities." But once the strategic session is done, they throw away all the ideas. This may not sound intuitive, but his company is a thriving, sustainable one.

The Dan Ward said...

And actually, I seem to recall the F-16 *was* designed to function as aerial artillery. It's early critics wrote it off as "a mud fighter," and disliked the fact that it was effective in an air-to-ground mission, as opposed to the Cadillac-esque F-15 Air Superiority fighter...

Mark said...

My observation on #2:

No one has said here that strategic planning is silly. Dan's original statement was that long-term strategic planning is silly.

If you are going to plan strategically for anything 20 years out, you probably would do best to focus your strategy on becoming more agile and able to react to inevitable (and unpredictable) change. In other words, don't strategize about the specific technology so much, instead strategize about how to be better prepared for technological curveballs, explosions, stagnations, etc. You will always be reacting to changes one way or another. How much of a struggle that is will depend not on a lucky guess from 20 years ago, but on the current ability to adapt.

On the flip side, if you are not doing some near-term strategic planning (say <5 years) then that is pretty silly, too.

The Dan Ward said...

I agree, Mark - short-term strategic planning is pretty important... and the best long-term strategic plan is to foster & develop flexibility and responsiveness.

Well said!