Paradox Unlocked

Executive summary

Leaders face demands that appear to conflict: cost and service, autonomy and control, central standards and local responsiveness, results now and capability for the future. Sometimes the choice is real. Often it reflects how the work has been arranged, how responsibilities are divided or how the problem has been defined.

Choosing a side and negotiating a compromise both have their place, but repeated compromise teaches people to play the game. The more valuable step is to change the work so that progress on one objective helps the other: removing the causes of customer calls improves service and reduces cost. Each such claim needs a mechanism, economics that work and evidence from a real test.

Top teams can create the trap themselves, when executives are labelled by the side they defend and the chief executive keeps deciding which function wins. Better thinking needs freedom from fear, shared purpose, time and a leader whose conduct makes the invitation credible. Decisions about AI deserve the same discipline: the opportunity comes from improving the work, not from counting automated contacts.

Start with the argument everyone already knows. Establish what each side is trying to protect, examine the cost of the current argument, test a specific change with an accountable leader and build what works into how the business runs. Where the conflict cannot be changed, make the choice and explain it.

Imagine the annual planning meeting.

The finance director wants to reduce costs. The commercial director wants to improve customer service. Both have compelling presentations. Both can explain why their proposal is essential to the strategy. Both suspect that the other has failed to understand how the business works.

The chief executive listens patiently, asks some sensible questions and eventually negotiates a compromise. Customer service gets some investment. Finance gets some savings. Everyone leaves with something, including a few new reasons to distrust their colleagues.

The meeting has produced a decision. Whether it has improved the business is another matter.

This is a familiar leadership problem. Organisations need growth and discipline, speed and rigour, individual accountability and collective effort. They need to deliver this year’s results while building the capabilities that will produce next year’s. Leaders spend much of their working lives dealing with demands that appear to pull in opposite directions.

Sometimes the choice is real. Money spent on one project cannot simultaneously fund another. A business running out of cash has an immediate problem that requires immediate action.

But sometimes the choice reflects how the organisation has arranged its work, divided its responsibilities or defined the problem. Accept those arrangements as fixed and the argument can continue indefinitely. Change them and a different answer becomes possible.

Better service can reduce costs. Greater autonomy can improve control. More rigorous preparation can accelerate execution.

Each of these statements needs a mechanism behind it. Without one, asking people to achieve both objectives simply gives them a more demanding version of the same problem.

In Accelerating Performance, we devoted a chapter to “Dissolving Paradox”.[1] Its central argument was that leaders can create substantial value by finding ways for apparently conflicting objectives to reinforce each other. We called the change that makes this possible the strategic unlock.

The phrase sounds grand. The work is usually quite practical. It might involve redesigning a process, developing a team’s capability, changing a measure or removing an internal boundary.

The consequences can be considerable.

Better service can reduce costs. Greater autonomy can improve control.

The question determines the answer

Return to the planning meeting. The question on the table is: how much customer service can we afford?

That question contains an assumption. It assumes that better service necessarily requires more expenditure. Given the way the business currently operates, that might be true. Customers call because deliveries are late, invoices are wrong or nobody can tell them what is happening. More calls require more people. More people increase costs.

But why are customers calling in the first place?

If the organisation removes the causes of those calls, customers get a better experience and the business has less work to do. Fewer errors mean fewer complaints, fewer repeat contacts and fewer refunds. Staff spend less time repairing yesterday’s mistakes.

The conversation changes. The team can now investigate which failures create the greatest combined burden for customers and the business, and what it would take to prevent them.

The original question encouraged an argument over a budget. The new question encourages the design of a better operation.

This is the practical importance of paradoxical thinking. It challenges the assumptions that make two desirable outcomes appear incompatible.

It also changes who needs to be involved. A discussion about the service budget may be dominated by finance and the contact centre. A discussion about preventing customer problems may require sales, operations, product design and technology. The source of the cost may sit several departments away from the budget in which it appears.

A surprising amount of organisational difficulty survives because nobody has responsibility for the whole problem.

Four ways of thinking

In the book, we described four levels of thinking. These are approaches to a decision. A capable leader needs access to all four and the judgement to know when each is useful.

At level one, we choose.

We can reduce costs or improve service. One objective wins. One loses.

There are circumstances in which this is exactly right. An urgent decision may require a clear instruction. An unattractive investment may need to be rejected. A business cannot pursue every opportunity.

The difficulty comes when choosing becomes the default response to every tension. If each objective has an executive champion, deciding which objective wins also decides which executive wins. The discussion becomes personal. People defend their position, recruit allies and rehearse why the other side is wrong.

The organisation may achieve clarity while losing cooperation.

At level two, we compromise.

We accept that both objectives matter and look for a tolerable settlement. Service gets half the requested investment. Finance gets half the proposed savings.

Compromise can be useful. It can secure agreement and allow work to proceed. But a settlement between executives is not necessarily a sound design for the business. Half the investment might be insufficient to improve service, while half the savings might leave the underlying economics untouched.

Repeated compromise also teaches people how to play the game. If proposals are routinely cut in half, asking for twice what you need becomes a sensible negotiating strategy. The planning process becomes more elaborate and less informative.

At level three, we find a clever way to accommodate both objectives.

Perhaps we offer different service levels to different customer groups. Perhaps we share resources or sequence investments. We work harder at the problem and discover possibilities that the original argument overlooked.

This is often valuable. Many good business decisions belong here.

The insight can be simple. Implementing it rarely is.

But clever solutions can accumulate complications. Different service promises need different processes. Exceptions need approval. Customer categories need updating. Someone has to maintain the spreadsheet that explains how it all works.

The arrangement may improve performance, but it can also create a growing burden of coordination. The leadership team becomes increasingly skilled at managing a difficulty that remains firmly in place.

At level four, we change the conditions that create the conflict.

In the service example, we investigate how getting the work right first time could improve the customer experience and reduce the cost of delivery. The ambition becomes a more reliable operation in which better service removes avoidable work.

This is the strategic unlock.

It requires more than a different sentence in the strategy document. Someone must identify the causes of failure, redesign the work and build the necessary capability. There may be investment before there are savings. Some responsibilities may need to move. Some measures may need to disappear.

The insight can be simple. Implementing it rarely is.

The economics must work

Toyota provides a useful illustration of the underlying logic.[2] Its production system includes jidoka: detecting an abnormality and stopping the equipment or line so that defective work does not continue through the process. Toyota describes this as a way to build quality into production and prevent problems from recurring. It sits within a system intended to deliver high quality, low cost and shorter lead times.

Stopping work can initially look like the enemy of productivity. Viewed across the whole process, however, preventing defects can reduce the subsequent burden of inspection, repair and wasted effort. The local interruption serves the wider flow of work.

That is the kind of mechanism leaders should look for.

The same reasoning can apply in a service business. An accurate invoice can improve the customer’s experience while reducing disputes and the work required to collect payment. Clearer product information can help a customer make a better choice while reducing returns. A more reliable delivery promise can reduce both customer frustration and the cost of handling exceptions.

These are possibilities to investigate. Their value depends on the facts.

A proposed improvement should therefore explain where the benefit will arise, what it will cost to achieve and how quickly it can be realised. Removing work creates capacity. Turning that capacity into a financial benefit may require further decisions about staffing, growth or the use of time.

Leaders should also check whether the apparent improvement merely moves the burden elsewhere. A shorter call is of little value if the customer has to call again. A lower departmental cost is unhelpful if another department inherits the work. A more convenient digital process for the company can still be a deeply inconvenient process for its customers.

The test is whether the whole activity performs better.

A surprising amount of organisational difficulty survives because nobody has responsibility for the whole problem.

There is evidence, with limits

The wider research on paradox offers support for taking these tensions seriously.

In a study published in 2018, Ella Miron-Spektor and colleagues examined a “paradox mindset”: the extent to which people accept tensions and find energy in working through them.[3] Their field study involved 135 product specialists in one large US company, with supervisor assessments of performance and innovation available for 91. Among those with a stronger paradox mindset, greater experienced tension was associated with better performance and innovation ratings.

This was an observational study in a particular setting. It does not establish that a workshop on paradox will cause higher profits. It does suggest that the way people approach competing demands matters to how effectively they work with them.

There is another qualification. Some tensions persist. Wendy Smith and Marianne Lewis’s work on paradox describes the continuing effort required to respond to interdependent demands.[4] An organisation will keep encountering questions about the present and the future, autonomy and coordination, stability and change.

Our practical interpretation is that a particular conflict can sometimes be dissolved by changing how the work operates, while the broader tension still requires attention. Improving today’s balance does not remove tomorrow’s leadership responsibility.

More autonomy can improve control

Consider a second familiar executive concern: how much freedom can we give people without losing control?

The answer often depends on what the leader means by control.

If control means personally approving everything, delegation necessarily reduces it. If control means achieving reliable outcomes, delegation may improve it.

Imagine a service manager who must approve every exception. The arrangement appears disciplined. Yet decisions wait in a queue, frontline colleagues stop exercising judgement and the manager receives increasingly compressed accounts of situations they have not witnessed.

The person with the authority has less information than the person facing the customer.

Becoming less necessary may be one of the most useful things you do.

A different arrangement would begin with clarity about the outcome, the boundaries within which people can act and the circumstances that require escalation. It would give colleagues the skills and information to make sound decisions, then use feedback to improve their judgement.

The manager’s work changes. Time previously spent approving routine decisions becomes available for coaching, reviewing patterns and addressing the causes of difficult cases.

The potential reinforcing effect is straightforward. Better capability makes greater autonomy possible. Greater autonomy gives people more opportunity to practise judgement and take responsibility. With effective feedback, that experience can strengthen capability further.

But the conditions matter. Delegating to an unprepared team without clear boundaries creates avoidable risk. Retaining every decision because the team is unprepared guarantees that it will remain dependent.

The leadership question becomes: what would we have to build for decisions to be taken closer to the information, with greater reliability?

That question leads towards investment in capability, clearer responsibilities and better feedback. It also asks something uncomfortable of the leader. Being necessary to every decision can feel rather good. Becoming less necessary may be one of the most useful things you do.

Common standards can enable local responsiveness

The argument between the centre and the business units is another promising place to look.

The centre wants consistency. Local leaders want freedom to respond to their markets. Each can produce examples of the damage caused by the other.

An organisation can spend years moving the boundary between them. It centralises after an expensive failure and decentralises after complaints about bureaucracy. New reporting lines appear. The underlying disagreement remains.

Part of the problem may be that “standardisation” and “local freedom” are being discussed as if each were a single thing.

They are not. A shared definition of customer data is different from a centrally dictated sales conversation. A common security requirement is different from a universal approval route for every commercial decision.

Consider an illustrative business with operations in several countries. A shared technical platform could remove the need for each country to build and maintain its own infrastructure. That would give local teams more capacity to develop offers suited to their customers. If useful local improvements could then be incorporated into the shared platform, local experimentation would also strengthen the common capability.

Standardisation would help local adaptation. Local adaptation would improve what the group shares.

Reaching that position requires precision. Leaders must decide which elements create value through consistency, where variation matters to customers and how learning will travel between the two.

It also requires a willingness to stop work. A common platform that leaves every local system intact may simply add another layer of cost. Local freedom that includes permission to ignore every common commitment can destroy the benefits of scale.

The strategic unlock lies in the design of the relationship between the centre and the businesses. Changing the organisation chart may be part of the answer, but the daily decisions still need to work.

Performance and health belong in the same conversation

A particularly costly apparent conflict sits at the heart of leadership: deliver results now or build the organisation for the future.

By organisational health, we mean the capacity to align around a direction, execute effectively and renew as circumstances change. These capabilities can sound attractive but distant when the current quarter is under pressure.

The discussion then becomes predictable. One group argues for results. Another argues for investment in people and capability. Each worries that the other has chosen the easier half of the leadership task.

A more useful starting point is the work that already determines performance.

Suppose an organisation struggles to turn customer demand into reliable delivery. Improving that process may require clearer priorities, better coordination between sales and operations, stronger supervision and more candid discussion of emerging problems.

Those changes build organisational capability while addressing a current commercial weakness. If they improve delivery, they may also reduce the volume of urgent interventions consuming management time. Some of that recovered capacity can support further improvement.

Performance and health can begin to reinforce each other through the same work.

This requires choices. Leaders cannot build every capability at once. They need to identify the few capabilities that matter most to the business’s performance and develop them through the delivery of real outcomes.

It also requires honesty about timing. Some investments will take longer to pay back than the business can currently afford. Others may be essential precisely because continuing to operate without them is so expensive.

Treating all capability investment as discretionary is poor judgement. Treating all of it as urgent is equally unhelpful.

AI deserves the same quality of thinking

A proposed use of AI can expose these assumptions very quickly.

Take a hypothetical customer service operation considering an AI assistant. The initial argument may be framed as automation versus personal service. One group wants lower costs. Another worries about customers losing access to someone who can help.

The leadership team could choose a side, negotiate a percentage of contacts to automate or create an elaborate division between customer groups.

It could also investigate a more consequential question: what prevents customers from getting their problem resolved correctly, and where could the technology help remove that difficulty?

An assistant might help colleagues retrieve information, prepare a response or identify recurring causes of contact. A redesigned process might make routine transactions easier while giving people more time for cases requiring judgement or empathy.

Those possibilities need to be tested. The right arrangement will depend on the task, the quality of the information and the consequences of error.

A business that measures only the number of automated contacts may congratulate itself while customers make repeated attempts to obtain an answer. Measures such as successful resolution, repeat contact, total handling cost and customer experience would provide a more useful test.

The commercial opportunity comes from improving the work. Buying the technology does not, by itself, settle the leadership question.

The commercial opportunity comes from improving the work.

The top team can create the trap

There is an uncomfortable human dimension to all this.

Executives rarely arrive at a meeting with a neutral attachment to their proposals. They bring responsibilities, loyalties and a history of what has worked for them. The finance director has been rewarded for discipline. The commercial director has been rewarded for growth. The risk director has good reasons to notice what others overlook.

Those perspectives are valuable. The difficulty begins when each becomes a claim to represent the whole truth.

Colleagues then acquire labels. Finance blocks growth. Sales makes promises that cannot be delivered. Operations resists change. The centre does not understand the market. Local businesses refuse to cooperate.

Once the labels settle, people spend less time investigating the issue and more time interpreting one another’s motives.

The chief executive can reinforce this pattern by repeatedly deciding which function wins. Team members learn to prepare a better case for their own position. The organisation becomes dependent on arbitration from the top.

A more demanding response is to require the team to work on the full problem.

The growth proposal must address the economics. The cost proposal must explain the effect on customers. The control proposal must account for the speed and practicality of delivery. Each executive must understand the legitimate concern behind a colleague’s argument.

This is where care and challenge can reinforce each other. Respect makes difficult concerns easier to raise. Serious examination of those concerns makes that respect credible.

The leader’s own behaviour matters enormously. If every unexpected objection produces irritation, the team will learn to offer fewer of them. If changing your mind is treated as weakness, people will defend yesterday’s position long after its usefulness has expired.

A good question, asked with real curiosity, can improve the quality of a meeting. The same words, delivered as an accusation, can shut it down.

Make level four possible

Our starting point is that everyone is capable of level-four thinking. It is what we can do at our very best: step back from a position, understand the legitimate needs on both sides and see how a different arrangement might serve them together. People move between the four levels. Experience and practice help us spend more time at our best.

Yet we often fall short of it. The same person who makes an imaginative connection in one conversation can become rigid and defensive in the next. A threatened budget, a public challenge or a fear of looking foolish can make protecting a position feel more urgent than improving the answer. Leaders need to pay attention to the conditions in which they ask people to think.

Freedom from fear is fundamental. People need to be able to question an assumption, admit that they do not know and put forward an unfinished idea without expecting humiliation or punishment. If the personal cost of being wrong is high, defending an established answer can be a perfectly rational response. The business then loses possibilities before anyone has examined them.

This is the practical meaning of psychological safety. Amy Edmondson’s research in 51 manufacturing teams found an association between psychological safety and learning behaviour.[5] It supports the importance of a climate in which people can take interpersonal risks; it does not establish that safety alone produces level-four thinking. High standards and serious examination of ideas remain essential.

A creative environment also gives people room to explore before they must defend a recommendation. Ask for several possible explanations. Bring in colleagues who see different parts of the work. Let people combine ideas, sketch a different process and ask what would have to be true for it to succeed. An early suggestion may be useful because of the next thought it prompts, even if it will never become the final answer.

The sequence matters. Generating possibilities and evaluating them both require rigour, but applying a final investment test to every half-formed thought can close the discussion too soon. Give an idea enough attention to understand its potential, then test it against customer needs, economics and practical constraints. Clear boundaries can make experimentation easier because people know where they have room to act.

Shared purpose gives that exploration a direction. Return the team to the outcome it is responsible for delivering together. Invite each person to explain the concern behind a colleague’s position before arguing for their own. Include people close to the customer and the process, and make room for quieter contributors. Rank is a poor guide to who can see the missing connection.

People also need enough time and attention to think. A difficult issue squeezed into the final ten minutes of an overloaded meeting is likely to receive a familiar answer. Where the decision allows it, pause, examine the assumptions and return with better questions. Leaders can do the same with their own reactions: notice the impulse to dismiss an objection and get curious about what it might reveal.

Above all, the leader must make these conditions credible through conduct. Acknowledge a useful challenge. Admit when new information changes your view. Ask what can be learned from a well-designed experiment that fails. Address ridicule, blame and the habit of claiming every good idea as your own. People will judge the invitation to think creatively by what happens to the first person who accepts it.

The responsibility is therefore shared. Each of us can practise better thinking. Leaders are responsible for making it easier for others to bring their best thinking to the work. That is how level four becomes a repeatable part of running the business.

People will judge the invitation to think creatively by what happens to the first person who accepts it.

Five things leaders can do

The approach becomes useful when it changes a live business decision. We suggest starting with one important tension that your team repeatedly fails to resolve.

First, establish what both sides are trying to protect. Move beyond the proposed solutions. “Centralise or decentralise?” describes two possible arrangements. It tells you little about the outcomes that matter. Perhaps the business needs reliable standards and faster responses to customers. State those needs clearly enough that everyone can recognise their legitimacy.

Check the constraints at the same time. Separate what is required by law, safety, cash or a genuine deadline from what is simply familiar. A long-standing convention can be surprisingly convincing when everyone calls it a constraint.

Second, examine the cost of the current argument. Look at delayed decisions, repeated work, customer problems and the management time consumed by escalation. Identify where one part of the organisation improves its own measure by making life harder for another.

This gives the discussion a commercial purpose. It also helps determine how much effort the problem deserves. A minor inconvenience should not become a six-month exercise in advanced thinking.

Third, ask how progress on one objective could help the other. What would have to change for better quality to reduce cost? For clearer standards to enable greater freedom? For more candid challenge to strengthen the team?

Require an explanation of the mechanism. Then investigate it with people who understand the work. The colleague dealing with the customer, operating the process or repairing the error may see something that the senior team’s presentations have concealed.

Fourth, test a specific change. Choose an area small enough to learn from but meaningful enough to matter. Establish a baseline, identify what will change and agree how you will assess both objectives. Include the cost of implementation and the work created elsewhere.

Name an accountable leader. Agree when to review the evidence and what would cause you to adapt or stop. An attractive idea should face the same discipline as any other investment.

Fifth, make the improvement part of how the business runs. A successful experiment will struggle to survive if budgets, incentives and approval processes continue to reward the old behaviour.

Change the relevant measures. Clarify responsibilities. Develop the capability required to sustain the new arrangement. Help others understand why it works so that they can apply the reasoning to their own decisions.

The objective is an organisation that can do more of this thinking for itself.

The objective is an organisation that can do more of this thinking for itself.

Know when to make the choice

There is a temptation to turn any useful idea into a universal answer. Paradoxical thinking is particularly vulnerable because the promise of achieving both objectives is so attractive.

Some demands remain incompatible within the available time and resources. Some proposed solutions cost more than they are worth. Some decisions are urgent enough that further exploration would be irresponsible.

A leader must be willing to make a choice, explain its consequences and accept what will be sacrificed.

There is also a difference between stretching a team and giving it an impossible assignment. Requiring higher output, lower cost and improved quality while refusing to change the process, provide capability or consider investment may simply transfer an unresolved contradiction to people with less power.

Calling that ambition does not make it achievable.

The discipline is to investigate whether the conflict can be changed, while retaining the judgement to act when it cannot.

The return is faster progress

In Speed Wins, we argued for reducing the time it takes an organisation to create value.[6] The way leaders handle competing demands has a direct bearing on that ambition.

Repeated arguments delay decisions. Political settlements create workarounds. Poorly designed controls turn leaders into bottlenecks. Customers experience the consequences long before the executive team agrees who owns the problem.

A useful strategic unlock changes that pattern. It can remove work, improve cooperation and release attention for the things that matter. Its value appears in outcomes: a problem resolved sooner, a more reliable service, a better use of cash or a team able to act with greater confidence.

That makes the next leadership meeting a sensible place to begin.

Look for the argument everyone already knows. The one with familiar positions, well-rehearsed objections and no satisfactory conclusion.

Then ask what would have to change for both legitimate objectives to become easier to achieve.

The answer may require a difficult choice. It may also reveal an opportunity that the argument has been hiding.

Sources and further reading

1. Price, C. and Toye, S. (2017). Accelerating Performance: How Organizations Can Mobilize, Execute, and Transform with Agility. Wiley. Chapter 16, “Dissolving Paradox”.

2. Toyota Motor Corporation. Toyota Production System. Accessed 5 October 2026.

3. Miron-Spektor, E., Ingram, A. E., Keller, J., Smith, W. K. and Lewis, M. W. (2018). Microfoundations of Organizational Paradox: The Problem Is How We Think About the Problem. Academy of Management Journal, 61(1), 26–45.

4. Smith, W. K. and Lewis, M. W. (2011). Toward a Theory of Paradox: A Dynamic Equilibrium Model of Organizing. Academy of Management Review, 36(2), 381–403.

5. Edmondson, A. (1999). Psychological Safety and Learning Behavior in Work Teams. Administrative Science Quarterly, 44(2), 350–383.

6. Price, C. and Toye, S. (2026). Speed Wins: How to Intelligently Increase Your Corporate Metabolic Rate. Accelerating Performance Consulting. Revised edition, September.