January 28th 2020

Forecasting and How to Improve It – with Margaret Heffernan

Why is our approach to forecasting so poor? Journalist, author and former CEO Margaret Heffernan on why even the best forecasters can only predict for 400 days, and on why greater emphasis is needed on experimentation and scenario planning.

Watch the full interview below. Members can also read the discussion paper on Future Insight, which Margaret co-authored with CRF Research Director, Gillian Pillans.


Margaret Heffernan [00:00:10] So I think there are really two fundamental ways in which the way we think about the future needs to change. And it starts with a recognition that we're very poor at forecasting. And this is not for lack of intelligence or a means or technology or anything like that. We live in a very highly complex environment where lots of different things have an impact in lots of different ways on lots of other different things. And actually keeping up with how that plays out in a highly non-linear fashion is extremely difficult. The best forecasters say they can go about 400 days out with a reasonable level of accuracy. That's because they're really good at it for 100 days. The rest of us, it's closer to 150 days. So what that means is that the huge amount of time and energy and effort that we invest in trying to get the perfect forecast is probably a waste of time. The second issue is that in the absence of forecasting, what we mostly are inclined to do is to refine our existing model. So we do what we've always done just a bit better or specifically a lot more efficiently. So better, faster, cheaper. And the problem is that if you're doing the wrong thing, doing it bigger, faster, cheaper is no solution. In fact, it may be exactly the thing that you shouldn't do. So I think our inability to forecast and our addiction to efficiency is a kind of overall governing principle means that we have, you know, around the world see institutions, commercial and otherwise really failed to to develop the adaptive and creative possibilities that you need to respond to what you really cannot see. Well, I think to be fair, I think some of those approaches were fit for purpose. Once upon a time, and because they were, we tend to cling to them because we think history repeats itself, even though no historian believes that. So we sort of think, well, if it worked last year, it'll work this year. That turns out to be wrong. So I think there are a whole bunch of things that we need to start thinking about. Instead, I think one is we need to be prepared to do a great deal more in the way of experimentation. If you inhabit a complex system, which almost all businesses do. Your ability to understand how it responds is quite limited. And what experiments do is they basically kind of test out the boundaries of the environment that you operate in. And I distinguish between experiments and pilots because pilots already heavily loaded and have a lot of political haft attached to them. Experiments are what you do when you don't know what to do and you think, OK, I can see life still probably going to get harder. I think we're going to be always under pressure to do more with less. So let's try stuff to see it in that domain. What's going to deliver results and what isn't? And out of that, I think, has come a way of thinking, which is quite close to open strategy, because you're drawing on the observations and questions of lots of people, every level of the organization. I don't think you do experiments by taking the senior leadership team away to a swanky hotel for the weekend. I think you have to get everyone's mind on the problem, on the problems. So I think that's one thing we have to develop by a critical capacity to in organizations, not as a life saving measure, not as some special fun, cute little exercise for the kids, but something we do all the time, everywhere. Anytime we can think of something that might make our capacity to adapt better, I think we have to continue to create stories about what the future of the organization might look like. The classic way of doing this is through scenario planning, but I think there are many flavours of scenario planning since it was developed by Shell. I think it's absolutely critical that it goes back to some of its traditional roots, which is to say hard data and soft data have to coexist within it. It has to be based on fact. You have to have contradictory scenarios and you have to involve a very large cross-section of people who have latitude to be critical and to say all the things which typically aren't allowed to be said in a corporate environment. So I think I think space for that is important. And I also think when Shell invented really scenario planning, one of the things that they never anticipated was that it would in itself change the culture because there's such a high priority placed on really radical challenge. And so it has this sort of double benefit, which is it definitely has a foresight capability, but it also develops culture on the move, which I think since culture is what provokes or sustains or undermines change is really fundamental. And the other thing of which of which comes into this in experiments also. But I think, you know, the third piece of this, I would say is it becomes increasingly important to see that companies are complex and blend of complicated and complex, and you need completely different mindsets to approach those different ways of being. So when you have, as most companies do, a lot of complicated processes, efficiency works really well in thinking through how can you make those bigger, faster, cheaper? It's catastrophic when you start addressing the complex. So I think you have to start thinking about the company as a sort of bi lingual beast, if you like, and which mindsets and tools are appropriate to each. And in my book, I talk about the importance of just in time thinking, which is about complicated and efficiency and just in case thinking, which is about having multiple solutions, multiple scenarios.


Margaret Heffernan [00:06:37] So that whatever actually does emerge, you have some preparedness for it. And I think preparedness is much more the order of the day than what we think of is conventional planning.