A valued former colleague once described a problem he kept encountering in his corporate innovation work:
“We need to align the different innovation teams with each other and with the core business. We need to find better tools to do that.”
At the time, the statement sounded completely reasonable. Different innovation teams pursue different opportunities, business units operate with different priorities, functions work under different constraints, and senior leaders often see only part of the overall system. Eventually, someone concludes that the organization has an alignment problem.
The response is usually familiar. Organizations organize alignment workshops, town halls, strategy days, cross-functional meetings, and portfolio sessions. They introduce shared terminology, common roadmaps, canvases, principles, and presentation decks. At the end, people may leave with a stronger sense that everyone is moving in the same direction.
My recent reading on group dynamics made me return to my colleague’s statement with a different question:
What if the alignment problem is not primarily caused by a lack of alignment?
What if what we call “misalignment” is simply the visible outcome of several very different mechanisms beneath the surface: deliberately different organizational roles, conflicting incentives, social identity, resource competition, hidden information, herding, polarization, or disagreements that have gradually shifted from the work itself to the people and groups doing the work?
If that is the case, an alignment event can easily address the symptom while leaving the cause untouched. Worse, in some situations it may strengthen the very dynamics that created the problem.
Before choosing an alignment tool, it may therefore be worth studying the causality behind the behavior we have labeled as misalignment.
Some misalignment is deliberately designed into innovation
The first complication appears when we look at organizational ambidexterity.
Charles O’Reilly and Michael Tushman describe the challenge established companies face when they try to exploit mature businesses while simultaneously exploring new ones [1]. Exploitation benefits from reliability, efficiency, predictability, and incremental improvement, while exploration requires greater tolerance for uncertainty, experimentation, changing assumptions, and business models that may not yet fit the operating logic of the existing organization.
This is one reason corporations create separate innovation units, venture teams, incubators, labs, or exploratory business units in the first place. Their separation is not necessarily an organizational defect. It is often part of the design.
An independent innovation team is therefore supposed to behave differently from the core business. It may work with more incomplete evidence, change direction more frequently, measure progress through learning rather than operating efficiency, and question assumptions on which the existing organization currently depends.
The core business, meanwhile, has a different job. It has customers to serve today, margins to protect, processes to operate, regulations to satisfy, people to coordinate, and commitments that cannot simply be abandoned whenever a new hypothesis appears.
These are not two versions of the same work.
They are different forms of work.
If we deliberately create that difference and later complain that the groups are insufficiently aligned, we may be diagnosing part of the intended organizational design as a behavioral failure.
That does not mean separation should become isolation. Ambidexterity also requires integration where strategy, resources, senior leadership, and eventual scaling connect exploratory work back to the larger enterprise [1]. The challenge is therefore not to maximize either separation or integration, but to understand what needs to remain different and what needs to become connected.
This creates an important first distinction:
Difference is not automatically misalignment.
Some differences should be protected because exploration and exploitation have different jobs. The more useful question is which differences are productive, which have become dysfunctional, and where coordination is actually required.
Already, “we need more alignment” becomes much less precise than it initially appears.
Structural separation can become psychological separation
Once a corporation separates innovation from the core structurally, however, something else can happen. The organizational boundary can gradually become a psychological one.
Social identity theory, associated particularly with Henri Tajfel and John Turner, explains how people categorize themselves and others through group membership and how these categories can influence perceptions and behavior between groups [2].
Corporations provide plenty of ready-made categories: innovation, digital, operations, strategy, ventures, finance, headquarters, business units, and the core.
At first, these labels simply describe different responsibilities. Over time, they can start to describe different kinds of people.
You can often hear the shift in everyday language.
People in an innovation unit may start saying that the core business is too conservative, too political, or incapable of understanding experimentation. People in the core may describe innovation teams as commercially naive, detached from customers, or protected from the consequences of their decisions.
Neither side needs to be completely wrong for the dynamic to become dangerous.
In fact, both sides often possess knowledge the organization needs.
The core business may understand customers, channels, regulation, operations, technical dependencies, and economics that innovation teams can underestimate. Innovation teams may be closer to emerging customer behavior, technologies, new business models, or changing competitive conditions that the established organization has strong incentives to interpret through the logic of the existing business.
The original purpose of structural separation was to preserve different capabilities.
The unintended consequence can be that useful difference slowly becomes identity.
The progression may begin with “we do different work,” move toward “we see things differently,” and eventually become “they are the problem.”
At that point, the discussion is no longer only about whether a claim is correct. The source of the claim begins to affect how the claim itself is judged.
Instead of asking, “Is this evidence convincing?”, people start asking, implicitly or explicitly, “Is this another thing innovation always says?” or “Is this just another objection from the core?”
Once that happens, the organization is dealing with more than an alignment problem. It is dealing with an intergroup problem.
The groups may also be competing for something real
Muzafer Sherif’s classic research on intergroup conflict adds another important mechanism. His work demonstrated how competition between groups could produce hostility, negative stereotypes, and social distance, while meaningful goals requiring cooperation across group boundaries could reduce those tensions [3].
A corporation is obviously not an experimental summer camp, but the causal question is highly relevant because innovation teams and established businesses frequently compete for scarce organizational resources.
They compete for capital, executive attention, people, technical capacity, customer access, strategic legitimacy, and sometimes influence over the future direction of the company itself.
Imagine an innovation team proposing a new business model that could eventually cannibalize an established revenue stream.
For the innovation team, the opportunity may represent the future.
For the existing business, the same opportunity may threaten current revenue, established expertise, authority, incentives, or organizational status.
That disagreement is no longer socially neutral.
If the tension is being generated by genuine resource competition, a workshop encouraging people to “understand each other better” may improve the tone of the conversation without changing the conditions that produced the conflict.
Sherif’s work is particularly useful here because the intervention was not simply more contact between groups. Cooperation improved around what he called superordinate goals: outcomes important to both groups that neither could achieve alone [3].
That suggests a more useful corporate innovation question than “How can we create more alignment?”
What important outcome requires capabilities from both the innovation unit and the core business, and cannot be achieved by either one alone?
That is a much more demanding question because it shifts attention from the meeting to the system that makes cooperation necessary.
There is another useful implication. The goal should probably not be to erase the identities of innovation and core business entirely. Research on common and dual identities suggests that intergroup bias can sometimes be reduced by strengthening a shared higher-level identity while still allowing meaningful subgroup identities to remain [4].
For an ambidextrous corporation, that makes intuitive sense. The objective is not to make exploration and exploitation identical. It is to let both remain different while preventing the difference from becoming tribal.
In other words:
Ambidexterity needs separation without tribalism and integration without conformity.
Strong alignment inside the groups can make alignment between them worse
This leads to another paradox.
Suppose an innovation unit becomes extremely aligned internally. Its members use the same terminology, attend similar conferences, read similar authors, and repeatedly reinforce the same beliefs about why the organization needs to change.
Meanwhile, the established business becomes equally coherent around operating experience, financial accountability, customer commitments, reliability, and the reasons execution matters.
Both groups may become more aligned internally while becoming less capable of understanding each other.
Research on group polarization matters here because group discussion does not necessarily pull people toward the middle. Under certain conditions, people who begin with a tendency in one direction can move toward a stronger version of that tendency after interacting with like-minded others. Research has examined both social comparison and persuasive argumentation as mechanisms behind this effect [5].
In corporate life, the shift may happen gradually.
Innovation moves from “the core may be too risk-averse” toward “the core cannot innovate.”
The core moves from “this initiative needs stronger evidence” toward “innovation teams never have evidence.”
Neither conclusion necessarily reflects stronger evidence. It may simply reflect repeated reinforcement among people who already see the issue in similar ways.
This creates a paradox that matters for any discussion of corporate innovation alignment:
The more internally aligned two groups become, the less aligned the organization as a whole may become.
The problem is therefore not simply the amount of alignment.
The better questions are:
Aligned around what? With whom? And at what point in the decision?
Groups can possess the information they need and still fail to use it
Another problem appears when we examine how groups exchange information.
Garold Stasser and William Titus showed that decision-making groups can disproportionately discuss information that members already share while failing to adequately surface information uniquely held by individuals [6]. This line of research became associated with the “hidden profile” problem: the group may collectively possess information pointing toward a better decision while still failing to discover that answer through discussion.
Consider a corporate innovation meeting.
Everyone knows the market is growing. Everyone has seen the opportunity deck. Everyone knows senior management supports the initiative. Everyone has heard the strategic narrative.
At the same time, one salesperson has repeatedly heard customers reject the proposed buying model. Someone in operations knows about a dependency that could undermine the economics. A researcher has evidence suggesting that the assumed customer problem occurs less frequently than expected. A technologist sees an emerging alternative that could make the proposed solution obsolete.
The shared information has an advantage: everyone can participate in it.
The contradictory information is distributed.
Now someone says:
“We need to leave this meeting aligned.”
That expectation can subtly affect what people contribute. Information that reinforces the shared story is easy to discuss. Information that challenges it may create friction, delay, or social risk.
This creates one of the most important dangers of premature alignment:
A group can become more aligned while becoming less informed.
The visible increase in consensus tells us very little about whether the quality of collective judgment has improved.
Interested in working together?
The dissenter may be doing useful cognitive work
Charlan Nemeth’s research on minority influence provides another reason not to equate disagreement with dysfunction. Her work showed that exposure to minority positions can stimulate more divergent thinking and encourage people to consider alternatives they may otherwise neglect [7].
The important point is not that dissenters are automatically right. A minority opinion is not better simply because fewer people hold it.
The value of dissent can lie in what it forces everyone else to do cognitively.
The person repeatedly asking, “What evidence do we actually have for this?” may slow the meeting down. The colleague saying, “I don’t think customers care enough” may prevent a clean consensus. Someone questioning the dominant strategy may be described as difficult or insufficiently aligned.
Yet removing that friction may also remove one of the mechanisms through which hidden assumptions become visible.
The better response is not to romanticize disagreement but to make it useful.
If one group believes customers will switch and another does not, the immediate objective should not be to decide which group sounds more convincing. It should be to identify what each side believes, what assumptions create the disagreement, and what evidence could distinguish between the competing explanations.
Disagreement can therefore become raw material for discovery rather than something that must immediately disappear.
But not every disagreement is the same
This is where I find another distinction useful.
A disagreement can remain about an object, or it can shift toward the subject.
By object, I mean the work itself: the opportunity, customer problem, assumption, evidence, business model, experiment, resource allocation, strategy, or decision.
By subject, I mean the person, function, team, or organization holding the position.
This is my practical language, but it maps closely to an established distinction in organizational psychology between task conflict and relationship conflict. Karen Jehn’s work distinguished disagreement about the content of the work from interpersonal incompatibility and tension [12]. Later meta-analytic work by Carsten De Dreu and Laurie Weingart also cautioned against the simplistic idea that task conflict is automatically beneficial; relationship conflict is particularly damaging, and even task conflict can undermine performance under many conditions [13].
That nuance matters because the goal is not to maximize disagreement.
It is to keep disagreements about the work from unnecessarily becoming judgments about the people doing the work.
Imagine two statements:
“They optimize an existing operating system.”
and:
“They are resistant to change.”
The first statement locates the difference primarily in the object: the role, incentives, responsibilities, and operating conditions under which the core business works.
The second relocates the explanation to the subject: they are the problem.
The same thing happens in the other direction.
Compare:
“They are working with assumptions that have not yet been validated.”
with:
“They live in fantasy land.”
The first statement can still be investigated. Which assumptions remain uncertain? What evidence exists? What would reduce that uncertainty?
The second gives us very little to test. It explains the disagreement through the supposed character of the other group.
That transition matters because object disagreements can still be examined through evidence. Subject disagreements are much more likely to trigger defensiveness, identity protection, attribution, and conflict.
The progression may look like this:
We disagree about the work
→ We disagree because they see the work differently
→ We disagree because they are different
→ They are the problem.
Once a corporation reaches the last stage, another alignment workshop may achieve very little because the participants are no longer only discussing competing views of an opportunity. They are also defending judgments about the people and groups holding those views.
A useful diagnostic question is:
If I removed the names of the people, teams, and functions from this disagreement, would there still be a substantive issue we could examine and test?
If the answer is yes, bring the conversation back to the object: the assumption, evidence, constraint, trade-off, or decision.
If the answer increasingly revolves around who the other people supposedly are, the organization may be dealing with an identity or relationship conflict rather than a simple difference in professional judgment.
Apparent alignment can also be herding
There is another route to consensus that matters particularly in hierarchical organizations.
People sometimes converge not because they independently reached the same conclusion, but because they observed other people reaching it first.
Bikhchandani, Hirshleifer, and Welch formalized this mechanism through informational cascades. Under certain conditions, people may rationally follow the visible behavior of earlier actors even when their private information points elsewhere [8].
Corporate innovation offers plenty of equivalent social signals:
“The CEO likes this.”
“Strategy has already selected it as a priority.”
“The steering committee approved it.”
“The other business units are joining.”
These statements may contain useful information, but they are not evidence that the underlying customer, technical, or business-model assumptions are correct.
They also change the social cost of disagreement.
Once several powerful people appear to favor one direction, later participants are no longer reacting only to the underlying evidence. They are also reacting to the visible positions already taken by others.
Eventually, the organization observes widespread agreement and interprets it as multiple independent confirmations.
But that can be badly misleading.
Ten people agreeing are not ten pieces of evidence if nine learned what to believe from the first.
This matters especially in steering committees, portfolio reviews, and strategy processes where senior positions become visible before independent judgments have been captured.
The organization may believe it has aggregated ten perspectives when it has actually replicated one perspective ten times.
Crowd wisdom can quietly turn into what I call mob wisdom
This connects to the familiar idea of the wisdom of crowds.
The useful part of a crowd is not simply that it contains many people. Independence matters.
Jan Lorenz and colleagues experimentally showed that social influence could reduce the diversity of estimates without improving collective accuracy, while also increasing confidence in the resulting convergence [9].
That creates a condition I informally think of as mob wisdom.
I do not use the term as an academic construct. I use it to describe a practical failure mode: people influence one another, positions converge, diversity disappears, and the resulting consensus is then interpreted as though it represented many independent observations.
The group looks wiser because it looks more certain.
But certainty and accuracy are not the same thing.
The scientific picture is also more interesting than simply saying that social influence is bad. Joshua Becker, Devon Brackbill, and Damon Centola found conditions under which decentralized social networks improved collective estimates [10], while Ethan Bernstein, Jesse Shore, and David Lazer found that intermittent interaction could improve average problem-solving performance while preserving stronger solutions better than constant interaction [11].
The lesson is therefore not to keep people apart.
It is to think much more carefully about the sequence of independence and interaction.
Instead of putting twenty people into a room, presenting the sponsor’s preferred answer, and then asking everyone to discuss it, capture some judgments first.
What do you currently believe?
What evidence supports that belief?
What evidence weakens it?
How confident are you?
What would change your mind?
Only then expose the differences.
Now the organization has something extremely valuable: a record of what people thought before the group started influencing what everyone thought.
This turns alignment into a process rather than a starting condition.
When disagreement turns into fighting, evidence becomes ammunition
There is one final transition that matters.
A disagreement can remain about propositions.
One person believes customers will switch. Another does not. They ask what evidence could discriminate between the two positions.
That can still generate learning.
But when the disagreement becomes attached to identity, the language changes.
“Innovation always exaggerates.”
“Finance blocks everything.”
“The business doesn’t understand the future.”
“These innovation people have no idea how the company works.”
Now evidence starts functioning less as information and more as ammunition.
A failed experiment confirms what the core always believed about innovation. A successful pilot confirms what innovation always believed about the legacy organization. Contradictory evidence is discounted because of where it came from.
The central question changes from:
“What should we learn?”
to:
“Which side is winning?”
At that point, another alignment event may even make things worse if it simply creates a new arena in which both sides defend their identities and histories.
One possible causal chain now becomes visible:
Structural separation → group identity → internal reinforcement → polarization → competition → subject-level conflict → defensive interpretation of evidence → intergroup fighting
At the end of this chain, management may observe the behavior and summarize the entire situation in three words:
“We need alignment.”
Perhaps.
But that label tells us almost nothing about what created the problem.
And without understanding the mechanism, it tells us very little about what intervention is likely to work.
Interested in working together?
Alignment is an outcome, not a diagnosis
None of this means organizations should celebrate permanent disagreement.
Innovation eventually requires coordinated action.
Capital has to be allocated. Products have to be built or stopped. Capabilities have to be committed. Ownership needs to become clear. At some point, keeping every possibility open becomes more expensive than choosing.
Alignment therefore matters.
The more useful question is when alignment becomes valuable and what exactly people need to align around.
During discovery, forcing agreement around an answer can suppress useful information, encourage herding, and eliminate alternative hypotheses before they have been adequately tested.
During commitment and execution, refusing to align around a decision can destroy coordination.
Perhaps the sequence therefore needs to look less like:
Discuss → align → execute
and more like:
Elicit independently → expose differences → understand the disagreement → test → decide → align
Good innovation alignment is not getting everybody to believe the same thing.
It is creating conditions in which independent knowledge can surface, competing interpretations can remain visible long enough to be examined, evidence can change positions, and the organization can converge when coordinated commitment becomes more valuable than continued plurality.
That brings me back to my former colleague.
He was right that different innovation teams and the core business need mechanisms for working together. What I would question today is whether the search should begin with alignment tools.
I would begin one level deeper:
What group dynamic is producing the behavior we are calling misalignment?
A deliberately different organizational role, resource competition, group identity, hidden information, herding, polarization, task disagreement, relationship conflict, and simple coordination failure can all look similar from a distance.
They require very different responses.
Before organizing another alignment town hall, study the causality.
The free section of INNOVATION& is offered freely, and you are welcome to keep reading it that way. If these articles help you make better decisions in your working life, please consider becoming a paid subscriber. Researching and writing them takes time, and paid subscriptions provide the income I need to continue. Your support also helps keep part of this work available to everyone.
Thanks for reading INNOVATION&! This post is public so feel free to share it.
Start paid learning and application section
Diagnose the Alignment Problem Before Trying to Solve It
If what we call misalignment can be produced by several different mechanisms, the practical challenge is to determine which mechanism is operating in a real innovation initiative.
The purpose of the following Alignment Causality Diagnostic is therefore not to create another alignment framework. It is to slow down the jump from observed disagreement to intervention.
Instead of starting with:
“How do we align these people?”
start with:
“What is producing the divergence, and does it actually need to disappear yet?”
