“Even in reasoning upon some subjects, it is a mistake to aim at an unattainable precision. It is better to be vaguely right than exactly wrong.”^[1]
…Carveth Read

Part 1 of this series described the results of a Santa Fe Institute competition where teams attempted to predict future results based on past data. This competition is particularly interesting because it looked at a best case scenario — all results were narrowly numerical with “true” answers fully known. Even under these ideal circumstances it was very difficult for teams to predict the future based on past data and no team did well with every situation. This should lead business leaders to be cautious about their ability to reliably predict future business results.

Unfortunately, the demand for accurate prediction remains a consistent error in almost every business — partly in response to investor or executive demands. Worse, traditional business practice continues to believe reliable prediction is not only possible — but achievable in every situation.

Yet because businesses are complex adaptive systems, their futures are not — and cannot be — predictable in any precise and accurate way. Worse, business desire to make the future predictable often damages long-term business health while increasing uncertainty.

Reminder: The Fundamentals of Complexity

Readers should recall that complexity science studies unusual behaviors which result among large numbers of parts (what are called complex adaptive systems) as they interact and adapt — results which are only known once they emerge. Further, such emergent results are non-linear results or gestalts — results which are more than and different from a sum of the parts which interacted leading to the results.

This definition, on its own, should begin our caution for prediction. Since we cannot know behaviors until they emerge, the study of parts cannot predict what will emerge within a complex situations. As we will discuss later in this post, while we may anticipate what might happen, we cannot predict it.

Every business is a complex adaptive system where the term “system” is used in its scientific original which allows looseness of boundaries and definitions. Further, the business  operates and interacts among many other complex adaptive systems such as markets, customer groups, and the economy. Why is this “looseness” important? Complex adaptive systems continually evolve and change as interactions change, some parts fall away, some parts change dramatically, and new ones continually come into play. While businesses want to believe systems can be clearly defined and self-contained, the true complex adaptive systems of business always involve, as they must, a great deal of ambiguity. (Readers can read more on complexity in this blog post.)

Some Ways Complexity Affects Prediction

The first, and most critical, reality of complexity is that it leads to gestalts — results which are more than and different from a sum of the parts. Thus, small efforts can lead to dramatic and outsized results while massive efforts can lead to little value. Such results are not predictable. Thus, when the first Starbucks store opened no one could predict a future market with $20B in revenue from selling coffee in stores. When Apple released its iPhone, vast numbers of pontificators predicted disaster — and no one predicted a future app market worth billions in annual revenue. By contrast, Google introduced its Glasses just as Facebook introduced its MetaVerse — through massive efforts which fizzled and ended to be of little value.

In my upcoming book readers will survey a set of what I call complex forces — accumulated effects of the many mechanisms and behaviors of complexity. These forces are inevitable and continual parts of doing business which make prediction impossible. Consider some of these forces:

• Emergence. The future within a business (a complex adaptive system) must emerge and cannot be known from a detailed study of the parts. Every business, then, faces emergent results affecting its most important activities.
• Connections. What emerges arises from and among a wide range of connections. We cannot know every part in advance much less its connections to other parts nor what passes along these connections.
• Instability. The healthiest companies continually work amidst — and thrive within — instability. Such instability is required if a company is to arrive at a future point in time with sufficient demand for its products.
• Adaptation. All business is done amid a continual adaptation of company, employees, market, technology, customers, and much more. Such adaptation prevents our ability to predict the future because as soon as any element adapts, all other elements begin adapting in response while the original element adapts as well — the great feedback loop of business health.
• Human Forces. Each human being is a complex system on their own as are all groups of humans — whether working together or simply interacting. Tremendous good emerges from these human forces — a good which AI cannot deliver. At the same time, there is also risk of disaster as when group think limits a company’s view of their reality.

Every business works within a continually adapting emergent environment where rigidly defined company systems might sometimes deliver excellent results — yet every rigidly defined business system eventually fails.

Other Complex Factors Affecting Prediction

Most Predictions Involve Numbers — But Numbers are a Reduction

Attempts to predict future business results almost always express these predictions with numbers — from costs to units sold, cash flow, and profit. Even an attempt to predict “action Y will be successful” implies some quantitative expression of “success.” Yet, numbers are reductions — abstractions which isolate specific factors away from connections in the real world. Thus, the Santa Fe competition circumstances were ideal for prediction. Rather than wanting know the whole future of a business, scientists attempted to predict only what numbers would come next in a time series free from connected complex realities. Teams found even this limited prediction to be extraordinarily unstable.

No company should believe numbers can be  reliably predicted apart from the whole world around them. At best, a company might lay out a large number of pre-conditions and expectations of interactions over time then suggest that IF the predicted future behaviors are true and no unexpected events occur, THEN X units will be sold. Even this, though, is not fully true as there are always unexpected shifts between when a prediction is made and when results are know.

Every Prediction Leads a Company to Adapt

The more companies claim predictions are accurate and precise, the more they adapt to those predictions — regardless of accuracy. Every prediction is, on its own, an intangible interacting part within the complex adaptive system which is a business. As soon as a prediction is shared the company adapts in response to that prediction — adaptations which often change the circumstance which were assumed in making the prediction. Suppose, then, a company predicts that “new product X will sell A units and generate $Y revenue in the first Z years.” Once broadcast within the company this prediction becomes sticky — hiding within company recesses when  the company adapts to it.

Thus, when factory managers are given a prediction they choose how to make a product based on the predicted volume of units. If actual results are far different, the factory has likely made decisions which are not quickly reversible leading the product to be made in the worst possible way for actual results. As a result, the presence of the prediction has made future profits worse than they might have been without those predictions.

Predictions often also have secondary complex influence. If the time between the initial prediction and results being known is long (perhaps 2 years in the case of consumer goods), company decisions made in the interim will be affected by the prediction without knowing its accuracy. When reality becomes known, then, companies have often adapted into situations where they cannot survive the reality of what happens or that reality leads them to face years of poor results.

The Dangerous Difference Between Looking Back in Time and Peering Forward in Time

In many ways, the belief that business futures are predictable comes from the way the passage of time warps the business understanding of projects and the realities of doing business. Consider the following illustration which covers a four year period with the company at the end of year two and entering year three.

The first half of the graph reflects a project-level (or company level) view of the past. Notice:

• The idealized path is far smoother than reality and the few bumps in the road are minimized. Why? We have already passed this way and know what we encountered.
• Note that only three critical decision points are recalled. We remember fewer important decisions than we encountered — and often are unaware of decisions (the grey shaded points) which had at least as much impact on our work as the ones we remember. Such is a reality of all organized human activity.
• Many important shifts or changes happened outside our control making companies likely to dismiss their importance.
• In the rearview mirror, we assume the context within which we did business was constant and fixed. Having lived through it, we come to believe the past was consistent and inevitable.
• Businesses must, to keep boards and investors happy, also claim that their current success is an “inevitable” result of its actions and choices.

More broadly, most businesses white-wash the realities of complexity from their view of the past because they don’t want to admit its effects and lack the ability to consider how it affected their work.

Turning to consider the reality which lies in front of the company notice a far different reality:

• The context within which we make our decision and do the work is continually changing — and we cannot predict or control these changes. As a result, all of our work is affected by realities we cannot predict but might be able to anticipate.
• Consider the decision points. The image accurately shows a very large number of bumps in the road as well as decision points. This is the reality we have forgotten in the past.
  • We cannot, today,  know all the future decision points nor how important each will be. Most we know only as they arrive or emerge.
  • Some future decisions will be made without managers and executives knowing that an important decision was made. Suppose a department manager chooses software for their team to use. Once in use, that software may end up adopted throughout the company because it is effective but through a process unaffected by traditional “logical” management theories.
  • Some decision points become clear only AFTER we work into the future on the path we unknowingly chose.
  • More of the decision paths lead to dead ends than lead to what we might consider successful outcomes.
  • Some decisions may lead us to wander far “off into the weeds.”
• Not shown here, our company, our project, and we, ourselves, are also continually changing and that change affects both which decision paths we take and which decision paths will lead to something we will ultimately call success.

Of course, traditional business planning lays out graphs like the above in tremendous detail. What they fail to see is the ACTUAL path of the company might relate to that graph for a short period but will eventually diverge so far that there will be virtually no relation between what was predicted and what happens.

“Anticipate” Amid the Inherent Lack of Certainty

The more aggressively a company attempts to precisely the future and deliver  on those predictions, the less that company will be able to survive. Such predictability, if achievable at all, arrives only through extraordinary limitations on a company’s field of action — limitations which prevent future health for the company.

A healthy business, then, will never have the certainty needed for predictions to be reliable — the nature of business reality prevents predictive clarity. Instead of predicting we should begin attempting to anticipate what might happen. Changing only this single word has tremendous value. Where predictions claim to be solid, firm, and precise, anticipation allows us to work amid uncertainty to succeed within a future which cannot be predicted.

Consider the change from predicting to anticipating. Asked to predict future sales, a sales team will begin building precise spreadsheets calculating precise numbers and spend great effort and considerable time doing this. Yet these numbers are always precisely wrong. A far better result is obtained through discussion with sales managers about what they anticipate might happen. Allowed to envision rough scenarios, the team is more articulate, accurate, informative, and reliable — revealing key insights about the business future which cannot be found in attempts to precisely predict.

An important part of “anticipating” is allowing multiple possible answers with no one expected to be “best.” By comparison, prediction narrows a business’ focus more than is reasonable amid complexity. If a company is asked for predictions, they need to determine whether those are critical.

Of course, most businesses mistakenly believe that making precise predictions makes results more likely. That’s wrong — the opposite might be true. After all, predictions are so brittle that any change in circumstances causes their usefulness to drop precipitously. Yet anticipating a variety of results leaves doors open to be prepared for what emerges.

Critically, the future path within any complex adaptive systems (e.g. companies) is always uncertain. So despite extensive flow charts predicting “what will happen” we cannot predict our path — much less know where it will lead us. This is a good thing in that it allows us to discover better paths or that our original path is no longer possible.

Whether through anticipation or any other method, businesses need to anticipate more and claim to predict far less.

[1] Carveth Read, Logic: Deductive and Inductive, (1920), page 352.

©2026 Doug Garnett — All Rights Reserved

My upcoming book, The Complexity Paradigm: Using System Science to Drive Business Success, will be available through Columbia Business School Publishing in January 2027 and can be pre-ordered today.

You can read more about my rather unusual background (math, aerospace, supercomputers, consumer goods & national TV ads) at www.Protonik.net. Through my company, Protonik LLC, I consult with companies as they design and bring to market new and innovative products while I have taught marketing, consumer behavior, and advertising at Portland State University since 2001. A member of the RetailWire.com Braintrust, I regularly engage discussions of retail challenges on topics of interest. Together with my podcast partner Shahin Khan, current issues in marketing and business are discussed on The Marketing (And Everything Else) Podcast — available on Google, Spotify, the OrionX website, and Apple Podcast.