Developing Intelligence for a Complex World


We’ve always assumed oracles of the future got their answers from breathing mountain air or from the gods looking down. The reality is the good oracles were canny oracles that did their homework first before delivering their fateful responses. They had access to great intelligence networks to help them, and most importantly before they gave a foresight response to a tough question they purposely disturbed the dynamics of the situation using quick probes to see what information would be revealed. We can’t look at a used car and expect to know how it will run; we have to kick the tires, turn on the engine, and give the car a test ride to have any sense of the future. Good oracles understood they weren’t passive participants in how the future played out; they were active, engaged ones. Whatever they said or did could influence the final outcome. Before delivering their foresight response, which was irreversible, they probed or stimulated the complex environment to see what possible behaviors could develop.

In modern day organizations, oracles have been replaced by effective intelligence practices that focus on getting “intelligence from the trenches,” conducting as much experimentation and testing in the market as possible, and learning quickly from emergent behaviors in the market. Google and Amazon have created amazing probing or active-intelligence capabilities that should enable them to remain very competitive for the foreseeable future.


Real-world environments are very messy. They evolve in nonlinear ways and are largely unpredictable. Even with a computer, we barely understand how the many forces interact and behave. We can have voluminous information about a situation, but it’s always very incomplete, often inaccurate, and very hard to integrate and analyze quickly. We generally have limited insight about the individual forces shaping the situation and the situation as a whole, and only understand in retrospect why things happened. We need intelligence practices that will help us better understand the situations we’re facing, that will coax out better foresights about future threats and opportunities, and that will enable us to act effectively in response to the changing environments.

Our intelligence needs start with understanding better the dynamics of complex environments. Real-world environments are complex because many interrelated forces interact unpredictably. Change in one type of force, social or economic or physical, inevitably affects the others, but the unfolding behavior of the system cannot be predicted by understanding the individual forces or any set of force interactions. Importantly, every stakeholder in a situation is a force, playing a role and acting on the other forces.

A complex environment is simultaneously dynamic and resilient. Resilience is the capacity of the environment to absorb disturbance, to undergo change, without crossing some threshold to chaos and a different dynamic. This capacity to undergo some change without a radical change in general dynamic is defined as the resilience of the system. The more resilient the system, the more anti-fragile (a term from Nassim Nicholas Taleb, the author of The Black Swan and Antifragile: Things That Gain from Disorder) it is.

The complex social-physical system becomes unstable and chaotic when changes in the interlinked forces result in thresholds being crossed. Shocks and disturbances to a system, such as from a natural disaster, technological disruption, etc., can push a system across thresholds into a different state or dynamic, often with unwelcome surprises. An accumulation of nonlinear changes can also push a situation over thresholds. Eventually the system reconfigures into the new dynamic and state with new thresholds.

Complex environments can be defined at many levels or scales. The highest level is the Earth system, but any real world issue, like the copper commodity market, the urbanization of Southern California, the development of renewable-energy sources in Europe, or migrations in Europe, can be defined as a complex situation.

Facing complex situations creates a number of strategy and intelligence challenges for organizations.

  • First, because real world situations are interactively complex and non-linear, they are difficult to explain, let alone predict some cause and effect. A relatively minor action like publicizing a foresight of the future can create disproportionately large effects. When Elon Musk makes a market prediction, he can have a large impact. However, the same prediction for the same issue at a later time may produce a different effect.
  • Second, each situation is unique and novel. Historical analogies can provide useful insights on individual aspects of the larger issue, but the differences among even similar situations can be profound and significant. The political goals at stake, the stakeholders involved, the cultural milieu, the histories, and other dynamics are unique.
  • Third, a complex situation can’t be known, only surrounded. The organization’s understanding of the issue depends on who’s involved, and each individual will see the relationships between the forces driving the situation and their importance differently.
  • Fourth, every description of the issue points in the direction of a set of foresights. The description puts blinders on what we see as foresight. For example, if one describes bankrupt commodity producers as the result of falling demand and lower commodity prices from a weak economy, the foresight of the future will be different than if we describe bankrupt commodity producers as the result of building too much supply capacity.
  • Fifth, any stakeholder’s action, including its intelligence activities, can disturb a situation with non-linear effects.
  • Six, real-world situations have unlimited possible outcomes; there’s no fixed set of possibilities. Also, there’s no way of knowing if many of the foresight possibilities have been identified and considered.
  • Seventh, foresight ideas for an issue are better or worse, not right or wrong. The suitability of a foresight and its perceived quality will depend upon how individual stakeholders have understood the situation and what constitutes success for them. The perceived quality of a foresight can change over time; yesterday’s foresight might appear good today, but disastrous tomorrow.
  • Eighth, every foresight for a real-world situation is a ‘one-shot operation.’ The interactive dynamics of a situation are continuously creating a new situation and cannot be undone. The consequences of change are effectively irreversible.
  • Ninth, real-world situations have no ‘stopping rule’. It is impossible to say conclusively that a situation has been resolved. Work will continue on an issue until strategic leaders judge the situation is “good enough,” or until stakeholder motivations, will, or resources have been diverted or exhausted.


Based on these challenges, we can surmise what were the secrets of the good oracles. How did they come up with good foresight for the tough questions of the day? There were seven things.

First, good oracles did their homework before coming up with a foresight response to a tough question posed to them.

Second, good oracles had intelligence teams to assist them. An oracle must develop a superior ability to identify signals of change from the external environment and see the new possibilities for all the players. The oracle needs a team to accomplish this and the team must specialize in watching complex situations—watching non-linear dynamics, emergent behaviors, etc., identifying the key uncertainties and the ranges of possible outcomes, spotting signals of change, gathering new data, managing probes in intense environments, and developing new insights on threats and opportunities. Oracle team persons need to have an entrepreneurial mindset to operate in those fluid situations, work with the open network, and communicate their insights upward. They must be able to recognize new patterns in a changing environment, know which types of relationships within the network are crucial at specific times, and mobilize relationships in order to accomplish objectives.

Third, good oracles developed a wide network of sources in the field, a network of players well beyond traditional players and boundaries.

Fourth, good oracles recognized they’re not outside observers in unfolding dramas, but active players in those dramas. Like every other player, an oracle action—particularly its foresight response—can shape the situation’s dynamics and possible outcomes.

Fifth, before settling on a foresight response, oracles first stretched their teams to identify the full range of possible dynamics in a situation and outcomes. The teams did this by focusing their intelligence activities on the big uncertainties, developing an understanding of the conditions from which opportunities or threats could emerge, and identifying the threshold boundaries—the tipping points—beyond which the possible outcomes and dynamics would change in chaotic ways. Very often the big uncertainties would be about the various players, who could emerge, how anyone might behave, etc. But there were many other possible big uncertainties—technology innovations, the success of new products and services, and local government rules.

Sixth, oracles extensively used probes. Instructive patterns can emerge from complex dynamics, if one can disturb or probe the situation and watch the effects. Probes can be a field test of a new product, an external-stakeholder interview, a publication of a blog, the posting of something to sell on eBay, etc. The objective was to coax out information about a major uncertainty, particularly about how key players might behave, so the oracle team could develop a better foresight response. Most probes will fail to produce anything, so oracles need a portfolio of them to create the opportunities for informative patterns to emerge.


Finally, good oracles made most of their money from retainer services because real-world situations rarely got resolved and clients wanted timely foresight updates. In fact, clients needed the oracles’ coherence in the midst of all the change, for seeing how the situational players were learning and adjusting to the changing dynamics. So good oracles developed processes for ongoing development of their network, watching of the dynamics of a situation, and probing the situation to stimulate new intelligence, if necessary.


In modern day organizations, oracles have been replaced by effective intelligence practices. Twenty years ago, Richard Pascale, former McKinsey consultant and Stanford Business School professor, described in a Sloan Management Review article, “Surfing the Edge of Chaos,” a set of strategic principles for organizations operating and competing in complex ecosystems and how Royal Dutch/Shell was attempting to apply those principles. One key principle was that in a world constantly evolving in ways one can’t predict, where one has a limited ability to understand the world and shape events and outcomes, organizations succeed best not by trying to control an unpredictable environment but by constantly disturbing it. Another principle was that decentralized organizational units were best positioned to develop the intelligence and insight for responding to the changing environment that was changing often in non-linear ways. At the time of the article, Royal Dutch/Shell was implementing a new management system that would rely on “intelligence from the trenches,” involve as much experimentation and testing in the market as possible, concentrate on rapid learning, and implement continuously adapting action plans. That new management system of Royal Dutch/Shell characterizes the approach many fast-growing corporations use today.

Probes are the techniques for generating intelligence from the trenches—by making small disturbances—and conducting market experiments and tests. They are the product tests, product announcements, market experiments, and interviews designed to get stakeholder responses. Probes are the means for resolving the uncertainties about what the stakeholders might do in the future. Will customers buy this new product idea? How might regulators respond to the new product or service? How might suppliers and competitors respond? Probes can be used to reveal emergent strategies of new entrants.

Online environments have totally changed how companies conduct probes. For example, the use of online surveys and tools like Survey Monkey has transformed how consumer research is done around the world. Corporations that compete in online industries and have access to millions of online users or customers are creating significant competitive advantages for themselves through their probes. They can run many probes, quickly, for little cost and are leveraging that capability to build their new products or services. Probes are a key for leveraging big data.

A major strategy of online companies is “ship and iterate.” This is essentially a strategy that leverages probing skills to commercialize a new product or service. The company doesn’t focus on getting a perfect first product introduced online but instead they initially make available online a close-enough product and then focus on iterating quickly to get improved versions into the hands of users. That first product shipment is in effect a big probe and generates a lot of useful information even if the product fails.

Google finds some of the most important data from the product-shipment probe is the negative feedback because it’s so motivating to the product development team. Google also believes in soft launching new products—i.e., only providing minimal marketing and public relations support with the initial launch, forcing the new product to gain momentum and succeed on its own.

Amazon believes speed matters with new products or services and, when there’s uncertainty about what might happen, it just tries something and takes advantage of the opportunities stimulated by doing something first in the marketplace. Being first also attracts to your product or service the critical segment of users and customers that is strategic and risk-taking—the innovators and generates for you the first feedback information from the marketplace that no one else will have. This bias for action is a characteristic of Amazon’s culture that is focused on continually trying to improve customer experiences.

For online and software companies, it’s easy to ship a new product or version. For hardware companies, it’s a little more difficult, but cost-effective probes can still be created. Online environments have enabled for hardware companies an array of new approaches and technologies for generating and testing product and market ideas fast, at low cost, and with not-much risk. For example, one can ship the design of the hardware, or one can create a virtual model that users can play with.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s