HANNAH BATES: Welcome to HBR On Strategy, case studies and conversations with the world’s top business and management experts, hand-selected to help you unlock new ways of doing business. Do you know the difference between effective forecasting and accurate forecasting? Technology forecaster Paul Saffo says accurate forecasting can be nearly impossible to do. But if you aim for effective forecasting, then at least you’re considering the full range of reasonable possibilities for the future.
In this episode you’ll learn Saffo’s six rules for effective forecasting – from visualizing future uncertainty, to discerning patterns in past data, and why it’s important to be your own worst critic of your forecasts. This episode originally aired on HBR IdeaCast in July 2007. And just a note – we recorded this by phone. While the audio quality is not great, the conversation is. I think you’ll enjoy it. Here it is.
ANNOUNCER: Hello and welcome to the HBR IdeaCast from Harvard Business Online. In this week’s episode, Cathy Olofson talks with Silicon Valley-based forecaster Paul Saffo about his article in the July/August issue of the magazine, “Six Rules for Effective Forecasting.”
CATHY OLOFSON: Hi, I’m Cathy Olofson from Harvard Business Review, and I’m here today with Paul Saffo. Paul is a veteran Silicon Valley technology forecaster and the author of “Six Rules for Effective Forecasting,” which appears in the July/August issue of Harvard Business Review. Thanks for joining us, Paul.
PAUL SAFFO: Pleasure to be here.
CATHY OLOFSON: So, in the article you say that executives should aim for effective forecasting, not necessarily accurate forecasting. I would have thought they were the same thing.
PAUL SAFFO: If you’re very lucky, they can be the same thing. But it is possible to be effective even when you’re not accurate. Accurate is very, very difficult and at times impossible, just because of the uncertainty of whatever situation is that you’re facing. If you head straight towards accurate forecasting, you may end up in a much worse situation by overlooking things that eventually come to pass. So, for me, effective forecasting means, rather than racing to a certain answer, you’re looking at the full range of uncertainty. Effective forecasting is all about understanding the uncertainty that lies ahead, and not rushing to a conclusion.
CATHY OLOFSON: So in the article you have six pretty interesting rules, and I thought we could go through those one by one. Principle number one is define a cone of uncertainty. What do you mean by that?
PAUL SAFFO: By the cone of uncertainty, it’s a common sense way of visualizing the uncertainty that lies ahead. As you’re standing in a given moment in time, if someone says, well what happens in two minutes? You can probably define a pretty narrow cone of uncertainty. As time goes outwards, the uncertainty becomes larger. You end up with a cone-shaped space. And what you really want to do when you make an effective forecast is to draw that cone in a way that it encompasses all of the reasonable possibilities that lie ahead. And keep that in mind as you’re working on your forecast.
CATHY OLOFSON: And rule number two is look for the S-curve.
PAUL SAFFO: Well it turns out everything interesting is shaped like an S-curve. Very few things unfold in a linear way. They instead unfold very, very slowly at the start and then eventually a critical point is reached, and everything starts changing rapidly. And it eventually settles and looks like an S-curve. The mother of all S-curves is the curve of Moore’s Law. That law described by Gordon Moore in the mid-1960s when he noted that the number of circuits we could put on a chip was doubling about every 12 to 18 months. And if you draw that out over time, it looks like an S-curve. So far it still looks like a hockey curve. But the bottom line is, it perfectly describes the whole digital revolution. It is why, for example, your wristwatch– if you still wear one– has more computing power on board than existed in the entire world before the early 1970s.
CATHY OLOFSON: It’s interesting here. You say that most ideas, even in Silicon Valley, take 20 years to become an overnight success.
PAUL SAFFO: Yeah. As a forecaster, one rule I live by is never mistake a clear view for a short distance. Just because it looks like something is really exciting and about to happen, chances are it’ll take a lot longer to arrive than you expect. And when it finally does arrive, it will unfold in unexpected ways. And it’s true again and again and again here in Silicon Valley. Quite simply, most ideas do take 20 years to become an overnight success. TV took 20 years from invention in the 1930s plus time out for a war before takeoff in the 1950s. Radio took 20 years from its first invention to the broadcast of station KDKA in 1921. The personal computer’s antecedent– it took the personal computer almost 20 years before it had takeoff. And of course the internet was almost exactly 20 years old when it took off in the year that Al Gore invented it, 1988.
CATHY OLOFSON: Rule number three: embrace the things that don’t fit. What do you mean by that?
PAUL SAFFO: Well as a forecaster, what I’ve developed is a finely-honed instinct for things that are strange. They don’t fit. They seem a little weird, or they’re intriguing but you don’t know why they matter. It turns out those are often like little Doppler whistles coming in from the future. And they’re very subtle indicators of something that may be a profound change around the corner. And so good forecasting often times is a little bit like bad research. You want to look for those subtle things, and say why does that bother me? Develop a forecaster’s intuition about which ones might be important.
CATHY OLOFSON: So, you say that indicators look more like oddball curiosities, which some people would say is what Second Life is. And yet it’s very popular.
PAUL SAFFO: Second Life is a marvelous example. The first oddball indicator of that that I experienced was a software environment called Habitat produced by Lucasfilm in late 1984. It was around in 1985, and disappeared. It was the first graphical multi-user environment that ended up being like Second Life, and it followed this classic pattern. It got my attention because it was the first graphical version of such a system. There had been text-based versions before. And I kept thinking, now what does this mean, and how soon could it happen? Sure enough, it followed that 20 years to become an overnight success pattern. Because Second Life was just the latest in a long line of programs that until then had been failures. One company after another tried to copy what Habitat did. Along comes Second Life in 2000. And guess what? Almost exactly 20 years, coincidentally, it takes off around 2005.
CATHY OLOFSON: Rule number four: hold strong opinions weakly.
PAUL SAFFO: Yes. Another rule I follow is strong opinions weakly held. I try to come to a conclusion as quickly as I can based on the information that’s available, and then I systematically attempt to dismantle my own conclusions. So come to conclusions quickly, but do not get attached to them. Forecasting proceeds as a sequence of failed forecasts. If you’re going to forecast, forecast often. Be your own worst critic of your own forecasts. Otherwise someone else will.
CATHY OLOFSON: Rule number five: look back twice as far as you look forward.
PAUL SAFFO: Well, you should look back twice as far as you’re looking forward in order to perceive the patterns that underlie the change you see. The simple fact is, even in periods of rapid change like today, the things that do not change are vastly greater in volume than the things that do change. And some of the truly unchanging things are the nature of human desires and human needs. The way they express themselves can be different, but there is a deep unchanging structure. And you want to look back twice as far as you’re looking forward in order to see if there’s a cyclic pattern. Mark Twain allegedly said– he didn’t really say it, but he should have– history doesn’t repeat itself, but sometimes it rhymes. And by looking back twice as far as you’re looking forward, you’re trying to pick those rhyming patterns and say, what’s the hidden constant? What’s the hidden pattern beneath the seeming novelties we’re facing?
CATHY OLOFSON: You say that one of the big mistakes that forecasters make is to use the past as support rather than illumination. And you talk about the kind of mistake that happened when Time Warner and AOL merged a few years ago.
PAUL SAFFO: What’s really important is to use history in the right way. Too often we use history the way a drunk uses a lamp post, for support rather than illumination. And it gets back to that being uncomfortable with uncertainty. Once you’re comfortable with uncertainty, then you can look back at history and pull the right lessons. A marvellous case of this in the political arena is the Iraq War. This administration and senior officials at the Pentagon studiously avoided looking at the Vietnam War by their own admission. Because they said, “We lost that war. There’s nothing to learn from it.” So, they cherry-picked history for the things that matched their preconceptions and desired results. That’s the road to ruin.
In the business world, the same pattern. Jerry Levin thought he could muscle his way into the digital age by acquiring AOL. By selectively looking at history. Not realizing that just as he bought AOL, there was a whole new change in the environment coming that would make AOL irrelevant. So, look back at history, but also look into the face of uncertainty. And don’t just cherry-pick the things that happen to fit your conclusion.
CATHY OLOFSON: You say that there are some situations where it’s impossible to make effective forecasts. And that’s rule number six: know when not to make a forecast. Can you tell us a little bit about that?
PAUL SAFFO: Sure. There are some times when uncertainty becomes so great you just have to pause, and say, “I’m going to wait a little bit until things settle down.” The moment of the fall of the Berlin Wall was a case in point, where the uncertainty there was just so great. You had to step back and say, “Let’s just see how things settle down a bit.” And sure enough, within about six months the fall of the Wall, the new world order became pretty clear, and you could start making a forecast.
CATHY OLOFSON: Do you think that say 20 years ago, anyone in the software industry could have seen the iPhone coming?
PAUL SAFFO: Well, devices like the iPhone have been talked about for a very long time. The term information appliance, and knowledge navigator– we’ve had visions in this space for a very long time. The question was, when would they become practical? And then what would the specifics be? So something like the iPhone generally was very predictable, and in fact there are some other things in the space. The wild card is the specifics.
A consumer electronic device like the iPhone lives or dies on the specifics of how it all comes together, not the general features. Again, to quote Mark Twain, he once said of writing, “The difference between the right word and the almost right word is the difference between the lightning and the lightning bug.” And that is so also true about features in consumer electronics.
So it would be easy to say, eventually these devices will arrive, and let’s keep looking for when they might break free. That would then put you in the right frame of mind to say when Steve Jobs said he was working on the iPhone, you could say, “Aha. Given his track record, given his sensitivity to interface design and building new models of interaction, this one looks like it is a candidate to be a launch to a new industry.”
CATHY OLOFSON: In Silicon Valley and in other sectors, the pace of innovation just seems breakneck these days. And I’m wondering if that makes forecasting harder, or easier, or it doesn’t make a difference?
PAUL SAFFO: My job is so easy because in fact, change is very slow. It seems like everything’s changing more quickly than ever today. What’s really going on is more is changing at once. And the acceleration effect we feel is the cross impact among multiple admittedly slowly changing curves. But when they intersect, things can suddenly take off like a rocket. So, the secret to understanding the pace of change and the outcome is to make sure you keep a broad peripheral vision. That you’re comfortable with uncertainty, and you don’t prematurely narrow that cone just to encompass what you hope would happen. You do those things, and it’s an easy business to be in.
CATHY OLOFSON: Well, that’s terrific. Thanks for joining us. Look forward to having you again soon.
PAUL SAFFO: My pleasure.
HANNAH BATES: That was technology forecaster Paul Saffo in conversation with Cathy Olofson on HBR IdeaCast. We’ll be back next Wednesday with another hand-picked conversation about business strategy from the Harvard Business Review. If you found this episode helpful, share it with your friends and colleagues, and follow our show on Apple Podcasts, Spotify, or wherever you get your podcasts. While you’re there, be sure to leave us a review.
And when you’re ready for more podcasts, articles, case studies, books, and videos with the world’s top business and management experts, find it all at HBR dot org. This episode was produced by Anne Saini and me, Hannah Bates. Ian Fox is our editor. Special thanks to Maureen Hoch, Adi Ignatius, Karen Player, Ramsey Khabbaz, Nicole Smith, Anne Bartholomew, and you – our listener. See you next week.
