Integrative Thinking 2.0: A User's Guide to Your Opposable Mind

ROTMAN Magazine

How to embrace opposing models and apply Integrative Thinking in four (not always easy) steps.

by Jennifer Riel and Roger Martin

IN THE EFFECTIVE EXECUTIVE, Peter Drucker writes at length about decision making, arguing that it is a central executive task. An effective decision-maker, he says, focuses on the most important decisions, works to achieve deep conceptual understanding and isn’t overly impressed by speed. But Drucker also points to a particular idiosyncrasy of effective decision-makers: “The understanding that underlies the right decision grows out of the clash of divergent opinions and out of serious consideration of the competing alternatives.”

Effective decision makers, Drucker says, disregard conventional wisdom about reaching consensus and instead work to create disagreement and dissention. As an example, he points to the man who turned General Motors into the largest company in the world:

“Alfred Sloan is reported to have said at a meeting of one of his top committees: ‘Gentlemen, I take it we are all in complete agreement on the decision here.’ Everyone around the

table nodded assent. ‘Then,’ continued Mr. Sloan, ‘I propose that we postpone further discussion of this matter until our next meeting to give ourselves time to develop disagreement and perhaps gain some understanding of what the decision is all about.’”

Sloan, Drucker says, “knew that the right decision demands adequate disagreement.” In other words, it is in the tension between competing ideas that we come to understand the true nature of a problem and start to see possibilities for a better answer.

This notion is at the very heart of Integrative Thinking. But it is also challenging to operationalize due to a key tenet of the human condition: conflict is uncomfortable and runs counter to our natural desire for certainty.We feel intuitively that opposing views are threatening to organizational harmony and that consensus should be our goal. No wonder, then, that when we’re faced with opposing options, we often discount one of them as simply wrong, and its proponents as either ‘misguided’ or ‘illintentioned’.

In fact, as Drucker hints, opposing models are only a problem when we choose to treat them as such. Sloan’s example offers another, more productive approach, which is to use conflicting ideas to truly understand the problem. We can dig deep into the opposing alternatives, and into the tension between them, to look for a better answer, treating opposing models as the raw materials—the building blocks—to create something new.

The question is how. Unfortunately, most successful integrative thinkers have, like Sloan, developed their integrative approach over a long career. Few could identify just how they explored opposing alternatives and created new answers. None have been explicitly taught what Integrative Thinking is or how to do it; they became integrative thinkers through trial and error, building idiosyncratic methods for integration, mainly implicitly and subconsciously.

Over the past seven years, we’ve developed a process to apply Integrative Thinking in a deliberate, conscious and directed way. Following this process won’t necessarily produce integrative solutions every time, but it will provide a higher probability of coming to a creative solution.

The Four Stages of Integrative Thinking

As indicated in Figure One, here are four key stages to the Integrative Thinking process. The first stage is to articulate the models — to tease out the opposing ideas. The second stage is to examine the models — thoughtfully, deeply and with as much affection as you can muster. These two stages are about deeply understanding the nature of the problem at hand — taking a deep dive into opposing models to understand what makes them work and what might be important to an integrative answer.

In stages three and four, we shift away from understanding existing models and towards generating new possibilities.  The third stage is to explore new possibilities — to ask, What kind of better answers might be possible? This is a rapid prototyping exercise — a generative challenge to explore a number of different resolutions, testing and refining as you go. The final stage is to assess the prototypes, in order to produce a comfortable level of confidence in your solution before moving ahead. After all, as Drucker said, unless a decision has been translated into action, “it is, at best, a good intention.”

At a glance, the process we have described appears to be linear. In practice, it isn’t quite so simple. Often, the real learning only comes with repetition, as your understanding shifts and deepens. Iteration is an inherent and important part of the process. For now, though, let’s assume a step-wise progression through the stages and explain each in turn.

STAGE 1: ARTICULATE THE MODELS

First, identify two extreme and opposing answers to the problem, turning an issue (e.g. how can we more effectively deliver training in a large multinational organization?) into a two-sided dilemma (centralize all training through HR vs. entirely decentralize training to the functions/regions.) In this stage, the goal is to make the solutions extreme expressions of a core idea.

Why two extreme ideas? First, it gives you a manageable place to start. Rather than having to work through a daunting range of ‘all possible answers’, it narrows the field to a manageable size. Second, by making the two options extreme alternatives, the starting options naturally subsume a large number of alternatives between them (i.e. many alternatives for the training issue fit on a continuum between ‘do it all centrally’ and ‘do it all locally’.)  If there is another very distinct option (like, ‘outsource all training to a global partner’), it can be included as a third opposing model. However, exploring a fundamental tension between two options tends to surface enough information to generate new possibilities.

It is important to sketch the two opposing ideas out to enough resolution that an observer could understand the essence of each model. This means taking the time to explain, in a few sentences, what each model would look like in practice (e.g. centralization = all development of training programs and content delivery is done at the head office, based on corporate priorities; decentralization = every unit is given the funds to develop or source their own local training, addressing whatever needs are most pressing for them).

Once the opposing models are clear, you explore each model in greater detail. To do so, ask who the key players are — the people most affected by the issue, who most need to be engaged by the answer. Often, these players will be customers, employees and the organization itself (as a proxy for shareholders.)  Alternatively, players could be different types of employees (e.g. HR, front-line employees and management), partners, government, etc. Exploring multiple perspectives helps to expand the salient features under consideration. For example, while an organization may care very much about how a model delivers profits or motivates employees, customers care most about the value a model delivers to them. Exploring the perspectives of multiple players helps to create a clearer picture of what really matters.

For each player, explore the benefits the model confers on them. This approach — focusing only on the supporting logic and not on the negatives — goes against conventional wisdom.  But focusing on the positive effects of the models, rather than looking generally at pros and cons, is intentional and important for three reasons:

• Citing negatives can easily shut down discussion of a given model; if a particular drawback seems insurmountable at the outset, it is hard to understand what might nonetheless be valuable in it;

• It is essential to understand the virtues, or what’s best about each model, so that valuable elements can be incorporated into a new integrative model; and

• If the models are truly opposing, the negatives of one model should be the positives of the other. For example, if we note that decentralization provides agility, it isn’t necessary to say that centralized models are often bureaucratic and slow.

When exploring the logic of each model, work in sequence and genuinely attempt to ‘fall in love’ with each model. Explore, as deeply as you can, what makes each model work well and what is valuable about it. For the moment, forget that any other models exist. Avoid judging or critiquing the models; the task isn’t to determine which is best, it is to consider rather than evaluate.  As an example, see page 8 for the logic of the centralization vs decentralization options.

STAGE 2: EXAMINE THE MODELS

Integrative Thinking is about leveraging the tension between models to create something new. So, once opposing models have been articulated separately, the next step is to look at the models together, explicitly holding them in tension. To do so, three sets of questions are helpful.

First, look across the models. How are they similar? For instance, ‘networking’ appears in both models as a benefit to employees.  In decentralization, the benefit is working with close peers; in centralization, it is meeting folks from across the company.  So, both models produce a networking benefit, but in very different ways. Start to consider how the benefit is produced differently in the two models, and how it might be produced in a new model. Then, consider the genuine points of tension between the models. For instance, centralization enables consistency across the organization, which is in tension with an ability to specifically address local needs in the decentralized model. It is hard to have both benefits at the same time; any  attempt to create a better answer should take this tension into account.

Second, after looking across the models, look within them.  Ask, What assumptions underlie each model? What are the crucial causal relationships? These questions are aimed at digging deeper into how the models work, where they break down, and how they might be understood differently. An assumption behind the centralized model, for instance, might be that employees from each region have more in common than not. An assumption of the decentralized model might be that individual needs are best understood and met locally. What if these assumptions

didn’t hold? How might you think about the problem then?

In terms of causal relationships, have a look at the critical outcomes of each model and how they are produced. What, really, is the relationship between autonomy and learning? Does local decision-making necessarily increase speed? And how does each model affect learning outcomes? By digging into the causal relationships, you can reconfigure the models thoughtfully and anticipate the effects of new models.

Third, look at the problem again with fresh eyes. Re-ask what problem you are trying to solve, recognizing that this may have shifted during the analysis. Is it really about the allocation of training dollars, or about finding the most effective path to a learning organization? Finally, ask which elements of each model you love and would want to keep in a new model (e.g. “I want both agility of execution and consistency of message.”) The purpose of this exploration is to find piece-parts of a potential solution.  Remember, there is no single right answer: the things you value in the models may be different from the things I value. But by identifying them, we can progress towards a range of possible ‘better worlds’.

STAGE 3: EXPLORE THE POSSIBILITIES

The third stage of the process signals a fundamental shift from analysis to creation. Once the models themselves, their respective benefits and their relationships are understood, you are ready to ask, How might they be integrated into a new and better answer?  One way to approach this stage is to reflect on your thinking and simply ask, How might I turn those elements I want to keep into a better model? (e.g. How might I create a training model that is both agile and consistent?)

This is not an easy task. It requires creativity, reflection, insight and some luck. Fortunately, when the answers aren’t forthcoming, or time is a pressing issue, the task can be made easier by exploring three guiding questions, which you can think of as a ‘search mechanism’ of sorts:

1. Under what conditions could one model actually generate one core benefit of the other?  Here, imagine taking one model and extending it in order to capture a single important benefit from the opposing model. Perhaps you liked the culture-reinforcing effect of centralization; how might you extend the decentralized model in such a way that it begins to reinforce something important about the company’s culture? Could a highly decentralized model, for instance, help to create and reinforce a culture of autonomy?

2. How could a new model be created using a small building block from each model?

Here, take one component from each model (such as ‘customizing content to context’ from decentralization and ‘economies of scale’ from centralization) and ask how they might be productively combined in new and interesting ways. Could an all-online model, for instance, keep costs low and yet be easily and quickly customized to a given context?

3. How might the problem be parsed in a new way, such that each model could be applied discretely?  In this case, explore how you might think differently about the problem, breaking it apart along an important fault line and applying each of your two opposing models to its distinct parts. For instance, you might break the training problem apart by task:  could the development of training and the delivery of training be divided in such a way as to get the best of both centralization and decentralization?

The goal of stage three is to create ‘prototype integrations’, so rather than censoring ideas at the outset, encourage wide and diverse suggestions. After generating a set of solutions, you’ll pare down the ideas as you explore them more deeply. Work to articulate what each prototype solution could be, and in the process, some solutions will move to the fore and some will fall away.

How will you know when you’re ready to move to the next stage? There is no algorithm to it. You will be ready to move on when you have made a genuine effort to work through the three questions, have generated several prototypes and have developed at least one answer that strikes you as having the potential to create more value than either of the initial opposing models.

STAGE 4: ASSESS THE PROTOTYPES

The final stage of the integrative process is to test your prototype solutions. Testing is crucial, because one of the most significant challenges for any new idea is a lack of data to prove that it will work. This presents a challenge: all organizations want innovation, but most feel safer in the status quo, so new initiatives are either quashed before they start (with a demand for pre-emptive proof) or sidelined when they fail to meet initial hopes and dreams. Given this dynamic, work to shorten the odds: test out the new ideas to create the data you need.

The simplest methodology for testing your prototypes is to share your ideas — as clearly and concretely as possible — with customers. Early on, get the prototype into the hands of real customers, and ask for feedback and suggestions, co-creating better prototypes together. For the training dilemma, you might generate a series of possible answers and storyboard the different experiences for users.

From there, employ a simple but essential question: What would have to be true for this integrative solution to be a good idea? It may seem counterintuitive — after all, isn’t the relevant question, What do we know to be true, relative to this idea? The problem with defining what we know to be true is that we get bogged down in arguments about different data sets and semantics about what constitutes ‘proof ’. Asking what would have to be true separates ideas from individuals and helps suspend judgment long enough to explore the ideas in full.

Let’s assume, for now, that we’re focusing on a single prototype answer (e.g. all curriculum development work is done centrally, as is regular training of geographically-distributed  trainers; those trainers then deliver and adapt content locally).  Prototype in mind, return to the same players you considered up front and ask, What would have to be true, relative to them, for the new solution to be a truly happy integrative answer? Reflect on what would have to be true about what each player wants and values to make this possibility a good one. For the training dilemma, for instance, it would have to be true that HR can effectively design all training programs without specialized local or functional knowledge. It would have to be true that the company will get better learning outcomes for the same or less investment.  And so on.

Once these conditions are captured, move on to identifying those conditions you are least confident actually hold true.  Remember, the daunting thing about new ideas is their lack of proof; knowing the specific aspects of the new idea that are most ‘worrisome’ allows you to set about generating data specifically directed at those worries. For each worrisome condition, a test can be designed to determine whether the condition holds.  So, you might build a pilot train-the-trainer program, or model out the costs of the different possibilities, or explore secondary research on learning outcomes under different delivery models.

In closing

As testing of your new model progresses and you gain confidence in the integrative solution, it is important to note that the challenge isn’t over. All solutions will eventually be made obsolete, and as a result, integrative thinkers tend to treat their solutions as provisional. As new models emerge in opposition to the integrative solution, the process begins anew. However, if applied thoughtfully, the process described herein will give you a fighting chance at resolving the wicked problems you face.

 

Jennifer Riel (MBA ’06) is associate director of the Desautels Centre for Integrative Thinking at the Rotman School of Management. Former Dean

Roger Martin is the Premier’s Research Chair in Competitiveness and Prosperity and Academic Director of the Martin Prosperity Institute at the Rotman School. He is the author of The Opposable Mind: How Successful Leaders Win Through Integrative Thinking (Harvard Business School Press, 2009).  His most recent book is Playing to Win: How Strategy Really Works (HBR Press, 2012), co-authored with A.G. Lafley.

Share this post