Organizational Diagnostic Models for Change Planning
Compare organizational diagnostic models used to assess readiness, map change impact, and guide planning for successful organizational change initiatives.
Why do so many diagnoses end with a confident fix for the wrong problem? Organizational diagnostic models exist to prevent exactly that, yet more than a dozen of them compete for attention, built on different logics, and picking one is already a diagnostic decision. Knowing what each was built to find is how a leadership team locates where effort will pay off.
What an Organizational Diagnostic Model Does
An organizational diagnostic model is a simplified map of an organization that names the variables worth examining and states how they are assumed to affect one another, so the evidence a diagnosis gathers is chosen by design instead of collected on hunches.
That choice is made before a single interview takes place, which is both the strength and the cost of working from a model. A sponsor who commissions a diagnosis without one usually gets a report that follows whoever spoke most forcefully: the head of sales blames the product, the product lead blames the forecast, and the consultant writes down both. A model decides in advance which evidence counts, so the diagnostic variables get examined whether or not anyone is complaining about them.
The case for that discipline has a number attached. Harvard Business School Online’s article Understanding the Congruence Model reports a survey in which 85 percent of C-suite executives agreed their organizations were bad at problem diagnosis, and 87 percent said those mistakes carried significant costs. Executives rarely lack data; what they lack is a frame that tells them which data bears on the problem in front of them.
A useful model also implies a performance standard. Once a model claims that structure, rewards and leadership should line up, any place where they do not becomes a measurable shortfall, and the diagnostician can judge where the organization falls short instead of only describing it. Used this way, the model works as an open-systems lens for locating where ineffectiveness originates, tracing an underperforming outcome back through the inputs and internal arrangements that produced it.
The full diagnostic process, from entry and contracting through data collection to feedback, is a discipline of its own within organizational diagnosis. The model is one input to that process: the part that decides where to look and what good looks like.
The Medical Diagnosis Analogy and Where It Breaks Down
The medical analogy holds that an organization, like a patient, shows symptoms that a trained diagnostician traces to an underlying condition before prescribing treatment. The Change Compass article How to Diagnose for Change builds its whole approach on this clinical sequence of assessment before intervention.
The analogy earns its place because it enforces order: examine first, prescribe second. Organizations that skip examination tend to buy the treatment they already wanted, whether a restructure, a leadership programme or a new tool, and then discover that the symptom outlives the cure.
Where it breaks down matters more than where it holds. A patient’s anatomy is fixed and documented; an organization has no standard anatomy, so the healthy state is a choice embedded in whichever model the diagnostician picked. Symptoms in medicine are measured by instruments; in organizations they are reported by people with stakes in the answer, so a finance team describing a “process problem” may be describing a turf dispute. And a patient does not change because a doctor asked questions, while an organization does: the act of interviewing people about what is broken raises expectations, surfaces grievances and shifts behaviour before any report is written.
The deepest break is about who holds the knowledge. In medicine the expert knows more than the patient about the condition. In an organization the people inside the system usually see their situation more clearly than any outsider, and the model’s job is to help them organise what they already know. A diagnostician who keeps the clinical posture to the end tends to produce a verdict nobody inside recognises, and an unrecognised verdict is rarely acted on.
Root Cause Analysis: What a Model Adds to an Open Search
Root cause analysis without a model is an open search that follows a chain of causes wherever it leads, and a model adds a fixed set of places to look so the search cannot stop at the first plausible culprit.
An open search, such as repeatedly asking why a problem occurred, has real strengths: it starts from the evidence and needs no theory. Its weakness is where it stops. Chains of “why” tend to halt at the first cause that is a person or a single process, because that answer feels actionable and nobody in the room wants to keep digging into areas they own.
Consider a business unit whose product launches slip quarter after quarter. An open search moves from late launches to late engineering estimates and ends at “engineering underestimates.” That answer is true and useless. A model with a rewards variable forces the next question: are estimates rewarded for optimism, because optimistic estimates win funding at the quarterly review? A model with a structure variable asks whether engineering is consulted only after commercial commitments have been made. The same evidence now points in three directions, and the team has to test each before choosing one.
A model also gives root cause analysis a stopping rule. The search ends when every variable the model names has been examined, instead of ending when the group runs out of patience. That makes findings comparable across units and repeatable a year later. The congruence framing, which treats mismatch between parts as the root cause, takes this furthest, and its limits show up as soon as several causes compete for attention.
The Major Diagnostic Models Compared in One Table
The major organizational diagnostic models span nearly three decades of published work, from Leavitt’s Diamond in 1965 to Burke-Litwin in 1992, and differ mainly in how many variables they name, which logic connects those variables, and what kind of problem each was built to examine.
A single table flattens differences that matter in practice. A four-variable model and a twelve-factor model can both claim to be “systemic,” yet one fits into a two-hour workshop and the other needs a validated survey instrument. The comparison table puts the candidates side by side so the shortlist can be made on one screen.
| Model | Year | Variables | Core logic | Best use |
|---|---|---|---|---|
| Leavitt’s Diamond (Lemon Learning) | 1965 | 4: people, tasks, structure, technology | Interdependence of four variables | Knock-on effects of a technology or process change |
| Weisbord Six-Box Model (Think Insights) | 1976 | 6: purposes, structure, relationships, rewards, leadership, helpful mechanisms | Box-by-box sweep for fit | First-pass diagnosis of everyday friction |
| McKinsey 7S Model (Prosci) | 1970s | 7: three hard, four soft elements | Alignment of all seven elements | Alignment check during a strategy shift |
| Understanding the Congruence Model (HBS Online) | , | 4: work, people, formal organization, culture | Congruence between components | Locating mismatch behind a recurring problem |
| Burke-Litwin Model (Lemon Learning) | 1992 | 12 factors | Causal ordering of factors | Planning a transformation |
| Galbraith Star Model | , | 5: strategy, structure, processes, rewards, people | Design levers reinforce each other | Redesigning an operating model |
| Lewin’s Force Field Analysis | , | 2 sets: driving and restraining forces | Balance of forces holds the status quo | Testing one proposed change |
Reading the Comparison Table
Origin Year and Author
Each row carries the date and the people behind it, because a model’s assumptions belong to the problems its authors were facing. Lemon Learning’s explainer on Leavitt’s Diamond dates the model to Harold J. Leavitt, an American industrial psychologist, in 1965. Marvin Weisbord published the six boxes in 1976 in Group and Organization Studies, and Prosci’s guide to the McKinsey 7S Model credits Robert Waterman and Tom Peters with developing it in the 1970s and splitting its elements into hard and soft groups. W. Warner Burke and George H. Litwin published their model in 1992 in the Journal of Management.
The year of origin predicts the gaps a model will have. Leavitt’s four variables include no box for culture or rewards, so a diagnosis run on the Diamond alone will not register an incentive problem unless someone files it under “people.” None of the dated models carries a variable for remote work, platform tools or informal digital networks, so a practitioner using a 1970s model today has to decide deliberately where those realities sit. Reading the date first tells a team which questions the model will never raise on its own.
Number of Variables
The variable count runs from four in Leavitt’s Diamond to twelve in Burke-Litwin, and each added variable buys coverage at the price of data collection. A four-variable model can be completed in a facilitated session with a leadership team and a whiteboard; a twelve-factor model usually needs a survey, a sample large enough to compare units, and an analyst to read the results.
The trade-off shows up in what the diagnosis can resolve. Fewer variables mean each one is broad, so “structure” in Leavitt’s Diamond swallows decision rights, reporting lines and coordination mechanisms in a single node. More variables split those apart, which helps when the problem sits in one of them and costs weeks when it does not. A team with a narrow, urgent question tends to get further with a small model applied quickly than with a large model applied once.
Core Logic and Best Use
The core-logic column sorts the models into four families: interdependence (Leavitt, Galbraith), fit (Weisbord, 7S, Nadler-Tushman), causal ordering (Burke-Litwin) and opposing forces (Lewin). Interdependence models say that changing one variable disturbs the others; fit models say performance depends on how well the parts match; the causal model says some factors drive others; force field analysis says the current state persists because pressures for and against change are balanced.
The best-use column is a starting hypothesis about where each logic pays off. Force field analysis suits a single, specific change because it needs a named target to weigh forces against; the Galbraith Star suits redesign because its points are levers a design team can pull. These entries are deliberately short: the actual choice depends on the problem in front of the team, and the families differ most sharply between fit and causation.
What Does Each Model Need Before It Can Be Run?
The table compares models by what they examine, but a shortlist also depends on what each one needs before a team can use it. Some rows are frameworks a facilitator can sketch from memory, while others only become useful once a survey, a sample and an analyst are in place.
Leavitt’s Diamond, the Weisbord boxes and force field analysis can be run as structured conversations: the model supplies the headings and the people in the room supply the evidence. Burke-Litwin sits at the other end, because twelve factors with a causal ordering are hard to test without survey data large enough to compare units. The 7S, Nadler-Tushman and the Galbraith Star fall in between and can be run either way, depending on how much the sponsor needs the findings to stand up to challenge. Asking what each candidate needs, alongside what it examines, keeps a team from choosing a model it cannot afford to run properly.
Older Models Still Cited: Likert System Analysis and Open Systems Theory
Likert System Analysis and open systems theory predate most of the table, yet both still appear in diagnostic reports because later models borrowed their core ideas.
Likert System Analysis places an organization on a spectrum of four management systems, from exploitative-autocratic through benevolent-autocratic and consultative to participative, using a questionnaire on how decisions, communication, motivation and control actually work. Its lasting contribution is the idea that management style can be surveyed and scored across a whole organization instead of inferred from interviews with a few senior people. Many later culture and climate surveys follow the same pattern of asking employees to rate how things work in practice.
Open systems theory is less a model than the ground the box models stand on. It describes an organization as importing energy, information and materials from its environment, transforming them through internal processes, exporting outputs, and depending on feedback from the environment to stay viable. Weisbord’s outer boundary, Burke-Litwin’s external environment factor and the inputs in congruence models all come from this view.
The reason it is still cited directly is its insistence on the boundary. A box model applied without the open-systems view tends to treat the organization as sealed, so a diagnosis concludes that structure is wrong when the real shift was a change in customers or regulation that the structure was built for. Practitioners who sketch inputs and outputs before opening any box are less likely to redesign an organization for an environment that no longer exists.
Fit Versus Causation: The Two Logics Behind Classic Models
Classic organizational diagnostic models follow one of two logics, fit or causation, and the split runs through congruence models built on mismatch, Burke-Litwin’s causal ordering of factors, and the formal and informal sides Weisbord asks diagnosticians to read in each box.
The two logics answer different questions, and an analyst who reaches for the wrong one gets a precise answer to a question nobody asked. Fit logic suits a stable organization with persistent friction; causal logic suits an organization about to change on purpose.
Congruence Models: Mismatch as the Diagnostic Signal
Congruence models treat a lack of organizational congruence, visible as mismatch between an organization’s parts, as the diagnostic signal, assuming performance follows from how well the parts match without saying which part drives the others.
Harvard Business School Online’s guide to the Understanding the Congruence Model credits organizational theorist David Nadler, with Michael Tushman and Charles O’Reilly, with creating it as a method of organizing company data to determine the root cause of a business issue, and describes it as the core component of root cause analysis. The McKinsey 7S Model and Weisbord’s boxes share the same assumption: a well-performing organization is one whose formal and informal organization, work and people are aligned.
That assumption carries a cost for the people acting on it. A fit model can return ten mismatches and still give no ranking of them, because a mismatch between structure and culture carries no information about whether structure should move toward culture or the reverse. Every mismatch looks equally urgent, and a leadership team with limited attention tends to fix the one it already wanted to fix, using the diagnosis as support.
The HBS guide offers a partial remedy: start with component tasks and their interdependencies, then design every other element around the work. That converts pure fit into a weak ordering, with the work as the anchor. It helps when the work itself is settled. When the work is the thing in question, because the business model is changing, the anchor moves, and a fit model has no way to say what should move with it.
Burke-Litwin’s Causal Ordering
Upstream Transformational Factors
W. Warner Burke and George Litwin’s 1992 paper, A Causal Model of Organizational Performance and Change, arranges twelve factors so that influence runs downward, and Lemon Learning’s guide to the Burke-Litwin Model places the external environment and leadership among the upstream factors whose changes flow down to work climate, motivation and, ultimately, individual and organizational performance. The upstream group is the external environment, mission and strategy, leadership, and organizational culture.
Burke and Litwin, arriving after a long run of large corporate transformations, built a model that predicts how a deliberate change will travel: Lemon Learning notes it was designed both to explain current performance and to anticipate how interventions ripple through the system. Changes to the upstream factors are transformational, altering the organization’s fundamental character, which is why a new strategy that leaves culture untouched tends to stall at the climate level. A practitioner can trace the expected path in advance and check, months later, whether it arrived.
Downstream Transactional Factors
The downstream factors are structure, management practices, systems, work-unit climate, task requirements and individual skills, individual needs and values, and motivation, all feeding individual and organizational performance. Changes here are transactional: they adjust the day-to-day exchanges between managers and staff, the procedures people follow and the climate of a unit, without touching mission or culture.
The causal ordering lets Burke-Litwin separate transformational levers from transactional ones, which a pure fit model cannot do. That separation stops a common waste: an organization trying to change its culture through a new performance-management system finds that the system is absorbed by the existing culture and turned into a box-ticking exercise. Read through the causal map, the mistake is visible in advance, because a downstream lever was asked to move an upstream factor.
Formal and Informal Sides in Weisbord’s Boxes
Weisbord asks the diagnostician to read every box twice, once for what the organization says should happen and once for what actually happens. Think Insights’ summary of the Weisbord Six-Box Model traces this to his 1976 paper, Organizational Diagnosis: Six Places to Look for Trouble with or without a Theory.
The formal dimension of a box is what documents, organization charts and policies prescribe; the informal dimension is what people do. Think Insights treats the formal-informal gap as the primary diagnostic target, because where the two diverge, formal changes alone tend to be neutralised by the informal system. Leadership sits at the centre of the six boxes with the job of maintaining fit between the formal and informal sides of the other five.
The informal side changes a diagnosis whenever the formal side looks healthy. A rewards box can show a documented bonus for cross-team collaboration while promotions consistently go to people who deliver alone; formally the box is fine, informally it teaches the opposite lesson. A diagnosis that reads only the formal side recommends better communication of the bonus scheme and changes nothing.
Weisbord’s premise of fit also faces outward. The boxes sit inside a boundary of inputs from the environment and outputs delivered back to it, and the fit between internal arrangements and that environment informs every box. In practice the pressures crossing that boundary come from customers, government regulators and unions, so a structure box that fits internally can still fail against a new regulatory demand. An organization that reads the informal side and the outward fit together catches problems that a formal, inward-looking sweep would leave in place for another planning cycle.
Culture-Focused Models: Competing Values and Denison
These two models put scores on the culture that structural models only name; when to start with one is covered under How to Choose a Diagnostic Model.
Scores carry their own risk: a number feels more certain than an interview, so a team may treat a culture profile as settled fact when it reflects one survey on one day. Prosci’s guide to the McKinsey 7S Model describes the soft elements, shared values among them, as less tangible and harder to measure, and the 7S supplies no instrument for measuring them. Culture models fill that gap, each with a different idea of what to measure against.
Competing Values: Four Culture Types and the Preferred-State Gap
Clan and Adhocracy
The Competing Values Framework grew out of Quinn and Rohrbaugh’s 1981 paper on organizational effectiveness, and Kim Cameron and Robert Quinn turned it into a culture diagnostic in Diagnosing and Changing Organizational Culture; Dave Snowden’s post The OODA loop & Cynefin (The Cynefin Company) traces both works. Umbrex’s summary of the Denison Organizational Culture Model describes the framework as mapping culture types along two axes, flexibility versus control and internal versus external focus. Clan culture sits on the flexible, internal side: collaboration, mentoring and loyalty hold the organization together. Adhocracy sits on the flexible, external side: experimentation, risk-taking and new products are what people admire.
The two flexible types matter for change work because they shape how a change will be received. A strong clan culture absorbs change through relationships, so a programme that bypasses trusted managers meets quiet resistance. A strong adhocracy welcomes novelty but loses interest in sustainment, so adoption peaks early and decays. Knowing which type dominates tells a change lead where the change is likely to stall.
Market and Hierarchy
Market culture sits on the controlled, external side, where competition, targets and winning define success; hierarchy culture sits on the controlled, internal side, where procedures, stability and efficiency hold things together. The framework treats all four types as present in every organization in different proportions, and the competing values are the tensions between opposite corners, such as clan against market.
The Organizational Culture Assessment Instrument turns this into a diagnosis by asking respondents to score the culture as it is now and the culture they would prefer, and the gap between the two profiles is the finding. A unit scoring high on hierarchy now and high on adhocracy as preferred is describing frustration with procedure, which points to specific levers: approval thresholds, experimentation budgets, how failure is treated. The gap also gives the people inside the system a shared picture they produced themselves, which tends to carry more weight than an outsider’s description.
Denison’s Four Traits
The Denison Organizational Culture Model diagnoses culture through four traits, involvement, consistency, adaptability and mission, each broken into three indices; Umbrex’s summary of the Denison Organizational Culture Model credits Daniel R. Denison and colleagues, with Denison Consulting later building the survey instrument and benchmarking database.
Involvement and consistency look inward, at whether people are engaged and whether values and agreements hold across units. Adaptability and mission look outward, at whether the organization reads its customers and whether it has a clear direction. The design links culture directly to performance outcomes, which makes it easier to present to an executive team that sees culture work as soft.
The comparison point differs from the Competing Values approach. The OCAI compares an organization with its own preferred state; Denison compares it with a benchmark of other organizations. That difference shapes what the finding means. An OCAI gap says “people here want something different,” which is a mandate for change. A low Denison score on adaptability says “organizations that perform well score higher,” which is an argument from external evidence.
A practitioner choosing between them should ask which argument the sponsor needs. A leadership team that doubts its people want change will be moved by an OCAI gap; one that doubts culture affects results will be moved by a benchmark. Picking the instrument that answers the wrong doubt produces a technically sound report that changes no minds.
Pairing a Culture Model With a Structural Model
A culture model works best alongside a structural one, because it shows what people value but says nothing about how work, structure and rewards line up to support or contradict those values.
The pairing earns its cost when the two readings disagree. A Nadler-Tushman pass may find formal structure and work well aligned, while an OCAI profile shows a strong preferred shift toward clan culture; the disagreement points to the informal side, where people feel the structure is efficient and isolating. Either model alone would have produced a partial story, and the structural model alone would have reported no problem at all.
The practical question is when a structural model’s culture box is description enough. A qualitative description works when the culture question is a single hypothesis that interviews can test, such as whether a recent leadership change has made people more cautious. A scored instrument becomes worth its cost when the decision depends on comparing units, tracking change over time, or settling a disagreement among leaders about whether culture is the problem at all. In those cases a description invites each leader to read their own view into it.
The order in which the two are run matters as well. Running the structural model first gives the culture survey a focus, so it can test specific suspicions instead of profiling everything. Running the culture instrument first risks a report full of interesting scores with no link to the levers the organization can pull, and that report tends to end up filed instead of used.
Integrative Models in Recent Research
Integrative organizational diagnostic models published between 2023 and 2026 mostly recombine mature frameworks, each adding one dimension the classic diagnostic models left thin, such as social networks, the realities of a small firm, or the unconscious dynamics that structural boxes cannot record.
That pattern raises a practical question for anyone tempted by a newly published model: does it add a variable the older frameworks missed, or does it repackage them for a particular context? The answer decides whether adopting it is a gain or a distraction.
Schiefloe’s Pentagon Model
Per Morten Schiefloe’s Pentagon Model sorts organizational issues into five interdependent categories: formal structure, technology, culture, interaction, and social relations and networks. His article Holistic Organizational Analysis: the Pentagon Model appeared in English in 2025 in the Journal of Organizational Sociology (De Gruyter).
Calling the Pentagon Model new would mislead a reader. The model refines Schiefloe’s earlier Norwegian work and builds openly on established organization theory; its first three categories map onto ground that Leavitt, Weisbord and Nadler-Tushman already cover. The article presents it as applicable at several levels, from individuals to whole systems and their wider ecology, which suits industrial settings where a single incident involves operators, contractors and regulators at once.
Its real addition is the fifth category. Classic box models mention informal relationships, but treat them as a quality of a box, as in Weisbord’s informal side. The Pentagon Model makes social relations and networks a category in their own right, alongside interaction as a separate category for how people actually coordinate. That gives informal ties the same diagnostic weight as formal structure.
The addition becomes measurable through Organizational Network Analysis, which maps who goes to whom for advice, information or decisions using surveys or communication metadata. Network data shows the brokers whose departure would cut two departments apart and the formal managers nobody consults. The classic models can only describe such ties; network analysis gives them numbers. A practitioner who adopts the Pentagon Model without collecting network data gains a label and little else.
A Holistic Model for Small and Medium Enterprises
Holistic models for micro, small and medium enterprises compress the classic boxes into a shape that fits a firm where the owner is often the strategy, the structure and the leadership all at once.
The classic models assume separated functions and management layers. In a thirty-person firm, a Weisbord diagnosis finds the same person at the centre of every box, and a Burke-Litwin survey has too few respondents to separate climate from individual opinion. Applying a large-firm model to a small firm tends to produce findings that are technically correct and practically empty: the structure box reports “centralised,” which everyone already knew.
Models sized for small enterprises change the variables instead of shrinking the method. They typically treat owner dependence as a variable in its own right, bring cash flow and customer concentration inside the diagnosis because they constrain every other choice, and read family and long-standing personal ties as part of the formal arrangement, since in a small firm the two are hard to separate. Data collection relies on conversations and observation over a few days instead of surveys.
The recombination logic holds here too. Little in a small-firm model is new theory; the contribution is a selection of classic variables weighted for the constraints of size. The test for adopting one is whether the firm’s presenting problem sits in owner dependence or resource limits. If it does, a sized model reaches it directly, and a large-firm model would spend most of its effort restating what the owner already knows.
Systems-Psychodynamic Diagnosis and the Unconscious Layer
Systems-psychodynamic diagnosis adds the invisible and unconscious aspects of an organization, its shared anxieties, defences and unspoken assumptions, to what structural models examine.
A structural model records that a division has been restructured three times in four years. A systems-psychodynamic reading asks what each restructure allowed people to avoid, for instance confronting a founder whose decisions nobody challenges, or acknowledging that the division’s core product is declining. Restructuring becomes a social defence: activity that relieves anxiety without touching its source. That reading explains why fit keeps failing to stick, something a box model cannot do because the cause is never written down.
The conceptual paper Organizational Diagnosis and Organizational Development Model: Integration of Psychoanalytic Determinants (ResearchGate) borrows the levels of dysfunction used in integrative medicine to help practitioners judge the nature and depth of organizational dysfunction, while preserving the collaborative, co-learning character of the development process. Depth is the point: some dysfunctions sit in procedures and respond to procedural fixes, while others sit in assumptions people cannot easily name.
The systems-psychodynamic approach has a real cost. It depends on skilled interpretation, its findings are hard to verify, and naming a defence can provoke the defence. Practitioners tend to use it as a second layer, brought in when a structural diagnosis has been acted on and the problem has returned. Used first, it risks a diagnosis that is insightful, unprovable and impossible for a sponsor to act on.
Strategic Diagnostic Models Versus Organizational Ones
Strategic diagnostic models judge whether a firm’s competitive position fits its market, while organizational diagnostic models judge whether the organization can execute the strategy it has, so confusing the two sends a diagnosis after the wrong kind of problem.
The symptoms look identical from the top. Falling margins, lost customers and missed targets can come from a strategy the market no longer rewards or from an organization that cannot deliver a sound one. A team that runs an organizational model on a strategy problem will find mismatches, fix them, and still lose.
Ansoff’s Strategic Success Hypothesis and Environmental Turbulence
Igor Ansoff’s Strategic Success Hypothesis holds that a firm performs best when its strategic aggressiveness and its capability responsiveness match the turbulence of its environment. A 2025 paper in Strategic Change (Wiley), The Efficacy of Igor Ansoff’s Strategic Success Hypothesis: A Comparative Analysis of Strategic Diagnostic Models, reads it as a dynamic, prescriptive and adaptive approach that aligns internal capabilities with external market conditions to guide strategic choices.
The dynamic label comes from the moving target. Environmental turbulence describes how fast and how unpredictably the environment changes, from repetitive markets where last year’s playbook still works to surprising ones where new competitors and technologies arrive without warning. As turbulence rises, the strategy and the capabilities that fit it change too, so the diagnosis is a reading of the gap between where the firm operates and where its environment now sits.
That gives the hypothesis a direct bridge to organizational diagnosis. Capability responsiveness includes management culture, structure and systems, the same territory organizational models cover. When the environment has moved to a higher turbulence level and the organization is still built for stability, the diagnosis is strategic and organizational at once, and the hypothesis says which way the organization must move.
The cost of the hypothesis is the data it needs. Rating turbulence and responsiveness involves judgement, and two teams can place the same market at different levels. The hypothesis works best as a structured argument among executives about how fast their environment is changing, and it loses value when treated as a precise score.
Static Tools: The Strategy Diamond and VRIO
The strategy diamond and the VRIO framework are static strategy tools: each assesses a firm’s position at one point in time, and the 2025 Strategic Change paper contrasts them with Ansoff’s dynamic approach.
The strategy diamond asks five questions about a strategy: which arenas the firm competes in, which vehicles it uses to get there, how it differentiates, how it stages its moves, and what economic logic produces its returns. It tests whether a strategy is coherent and complete. VRIO tests whether resources are valuable, rare, inimitable and organized to capture value, and a resource that passes all four tests is a source of sustained advantage.
Both tools ask strategy questions about market position. VRIO’s final test, whether the firm is organized to exploit a resource, looks like an organizational question, but it asks only whether organization stands in the way of exploiting a particular advantage. It does not diagnose how the organization works overall, and practitioners who treat a VRIO pass as an organizational diagnosis miss the culture, climate and informal side entirely.
A static reading serves well in a stable market: when turbulence is low, a snapshot of position stays valid long enough to act on, and these tools are quicker to apply than Ansoff’s framework. Their weakness shows when the environment shifts faster than the planning cycle: a VRIO assessment completed in January may describe an advantage a competitor has already copied by June.
Where the Two Families Meet
Strategy tools belong in an organizational diagnosis only where an organizational model, such as the strategy element of the 7S, shows that the strategy itself is in doubt.
The routing signal is a specific question: if every mismatch the organizational model found were fixed, would the firm then be executing a strategy the market rewards? If the answer is yes, the problem is execution, and the diagnosis stays inside the organizational model. If the answer is no, or nobody can answer it, the strategy element has failed, and a strategy tool such as VRIO or Ansoff’s hypothesis should enter before any redesign starts.
The routing matters most for incoming leaders, who face pressure to diagnose quickly. Harvard Business Review’s The New Leader’s Guide to Diagnosing the Business reports, citing Challenger, Gray & Christmas, that about 40 percent of CEOs who left their jobs in 2006 had lasted an average of just 1.8 years. Under that clock, a new leader is tempted to reorganise, because reorganisation is visible, and a reorganisation aimed at a strategy problem uses up the time the leader had.
The reverse error also occurs in practice. A team convinced that strategy is the issue commissions a market analysis, produces a better strategy, and hands it to an organization that could not execute the old one. Running a quick organizational pass alongside the strategy work keeps that from happening, because it shows whether the organization can carry any strategy before the new one is launched.
How to Choose a Diagnostic Model
Choosing among organizational diagnostic models starts from the presenting problem the client brought, then routes by the type of change and the size of the organization, because no diagnostic model is best in general and each fits a particular kind of trouble.
The presenting problem is often a symptom, and a model chosen to fit it can inherit its framing. A diagnostic model shows its users the variables it names and hides the rest, so diagnostic model selection is already a bet about where the cause lies. Three tests sort the candidates: whether its variables match the presenting problem, whether its assumed relationships can be checked against data the organization already holds, and whether the people who will act on the findings can understand and use what it produces. Small firms add a fourth consideration, size, which points toward the holistic small-enterprise models discussed earlier.
Matching the Model to the Problem
Transformational Change
A transformational change alters mission, strategy, leadership or culture, as in a merger, a shift in business model or a move from functional departments to product-aligned teams. For this kind of change a causal model such as Burke-Litwin fits, because the change begins in exactly the factors Burke-Litwin places at the top of its map.
A fit model struggles during transformation for a simple reason: everything is deliberately out of alignment while the change is underway. A bank moving to product teams will show mismatches between structure and systems, systems and skills, skills and rewards, and a fit model will report all of them as problems. The causal model separates the mismatches that are expected stages of the change from the ones that signal the change has stalled.
Everyday Friction
Everyday friction means recurring problems in a stable organization: handoffs that fail, decisions that take too long, units that work against each other. A fit model such as Weisbord or Nadler-Tushman suits it, because the question is where the parts have drifted apart while the organization stays broadly the same.
Friction problems are usually local and specific, and fit models handle them well because they can be run quickly with the people involved. A Weisbord sweep of a single department, done in a workshop, often surfaces an informal rule that contradicts a formal one within an hour. Bringing a twelve-factor causal survey to a handoff problem spends weeks of analysis on a question the team could have answered in an afternoon.
Values and Culture
When the presenting problem is phrased as behaviour, such as people not speaking up, silos, or resistance to anything new, a culture model such as the Competing Values Framework or Denison’s traits is the better starting point. Those problems sit in what people value and assume, and the culture instruments scored earlier give a team numbers to argue over.
The route matters because behavioural complaints tempt leaders toward structural fixes. A silo complaint answered with a reorganisation moves the silo walls without changing why people protect their territory. Starting with a culture instrument shows whether the silo reflects a market or hierarchy culture that rewards defending one’s own numbers, and that finding then tells the structural model where to look.
Choosing by Pattern: Bain’s Organizational Navigator
Bain & Company’s Organizational Navigator chooses levers by pattern: it links organizational outcomes to a small set of recurring patterns, so executives identify which pattern they fit before deciding where to invest.
Bain’s article Finding the Right Remedy for Poor Organizational Performance (Bain & Company), by Dan Schwartz, Ludovica Mottura and Julie Coffman, finds that six organizational outcomes enable strong business performance and that companies often exhibit one of five organizational patterns. It starts from a specific situation: the strategy is sound but performance has stalled, so the problem lies in how the organization orients around the strategy, executes against it, or both.
The article’s examples show why the pattern matters more than the symptom. A large technology service firm with strong execution misses growth and profitability targets because its senior team keeps changing operational objectives and metrics, leaving employees exhausted and confused; its problem is orientation. An industrial supplier with a strategy the organization already understands keeps holding multiday strategy offsites while execution against customer expectations stays weak; its problem is execution. Both firms underperform, and the remedies point in opposite directions.
Pattern-based selection suits executives who need a fast, defensible starting point. Its limit is that patterns are averages across many companies, and an organization that sits between two patterns can be pushed into the wrong one. The pattern works best as a hypothesis a team then tests with its own data before committing resources.
When to Step Back to Cynefin
The Cynefin Framework should come first when it is unclear whether any fixed diagnostic model suits the situation, because it classifies the situation itself before any model is applied.
Every fixed model assumes that cause and effect can be known in advance, which is what lets it state how its variables affect one another. In a complex situation, cause and effect become visible only in hindsight; in a chaotic one, there is no time to diagnose before acting. Applying a box model to either produces a confident map of relationships that do not hold.
Dave Snowden, writing on The OODA loop & Cynefin for The Cynefin Company, argues that Cynefin and the Competing Values model do different things, so overlaying one on the other makes little sense. That is the right way to use Cynefin alongside diagnostic models: as a prior step that decides whether a model’s assumptions apply at all.
In practice, a team that finds its situation complex shifts from diagnosis to probing: small, safe-to-fail experiments that reveal how the system responds, with a fixed model brought in later for the parts that turn out to be stable. A team that skips the step spends its diagnostic budget building a precise picture of a system that will have changed before the report is read.
Where Diagnostic Models Fall Short and How They Are Combined
Organizational diagnostic models fall short when they are treated as expert verdicts, because each one is an orientation device that captures a snapshot, and practitioners combine diagnostic models with feedback tools, team measures and task-level checklists to check one reading against another.
The limits are useful in their own right, because they tell a practitioner when a model has done its job and something else has to take over.
Orientation Devices, Not Verdicts
A diagnostic model works as an orientation device that focuses attention on where to look, and its output is a set of hypotheses for the people inside the organization to test before anyone acts on them.
The verdict posture fails in a predictable sequence. An outside team runs the model, writes a report, and presents findings to leaders who did not take part in producing them. The leaders defend their areas, question the method, and adopt the recommendations that cost least. A year later the presenting problem is still there, and the organization concludes that diagnosis does not work.
Every model also captures a snapshot. A diagnosis repeated after a change can disagree with the first one, and that disagreement is information: it shows what moved and what did not. Organizations that treat the first diagnosis as permanent truth miss the second reading, which is often the more useful one.
The practical signals that a model has reached its limit are visible in the data. Findings stop changing when the model is rerun. The same variable keeps lighting up, yet interventions aimed at it do not move it. Or the variation between units grows larger than the variation between the model’s variables, which means the model’s organization-wide view is averaging over the differences that matter. Any of these is the cue to hand over to a different tool instead of running the same model again.
Cross-Checking One Lens With Another
Practitioners combine a structural model with a culture model or with 360-degree feedback so that each lens checks the other, and the places where they disagree become the most informative findings.
A Weisbord pass based on leadership interviews may report the leadership box as healthy. A 360-degree feedback round in the same unit may show managers rated low on listening by their direct reports. The structural reading reflects what leaders believe about their own leadership; the feedback reflects what the people below them experience. Neither is wrong, and the gap between them locates the problem more precisely than either alone.
Team-level measures add a further cross-check. The Team Climate Inventory asks team members about shared vision, participative safety, task orientation and support for innovation; measures of psychological safety ask whether people believe they can raise problems without being punished. Both show variation inside units that an organization-wide model averages away. Two teams in the same division, under the same structure and rewards, can differ sharply in whether people report errors, and that difference explains more about performance than any box-level finding.
Model triangulation of this kind has a cost: more data, more analysis, and the risk of a report that lists every disagreement without resolving any. The discipline is to choose the second lens to test a specific doubt raised by the first, and to stop adding lenses once the doubt is settled.
Handing Over to Task-Level Tools: The PDC-HS
The Performance Diagnostic Checklist; Human Services takes over where organization-level models stop, diagnosing why specific people are not performing a specific task to standard.
A 2022 study in the Journal of Organizational Behavior Management, A preliminary investigation of procedural refinements to the performance diagnostic checklist – human services, describes the PDC-HS as a functional assessment tool used in Organizational Behavior Management to assess reasons for employee performance problems and inform intervention development. The study proposes refinements to the intervention-selection process, because two aspects of that process could undermine the tool’s objectivity and usefulness, particularly for practitioners with limited experience conducting the assessment.
The handover point is concrete in most settings. When an organization-level model has narrowed the problem to “care staff in two homes are not completing medication records correctly,” no box model can go further. The checklist asks whether staff were trained in the task, whether they know what is expected and are prompted at the right moment, whether the materials and processes make the task possible, and whether the consequences that follow support doing it correctly. Each answer points to a different intervention.
The two levels need each other. A task-level tool applied without an organization-level view can fix a symptom whose cause sits in rewards or structure, so the same failure reappears in another task. An organization-level model without a task-level follow-through produces findings that never reach the work.
How to Start Applying Organizational Diagnostic Models
The most reliable way to start applying organizational diagnostic models is to run one small, bounded diagnosis on a live problem in a single unit, using data the organization already holds, and to let the people in that unit interpret the findings before anyone outside it does.
The first step is to write the presenting problem in the words of the people who feel it, leaving the sponsor’s summary aside for now. “Releases slip every quarter and the team works weekends to recover” gives a diagnosis something to test; “we need a more agile culture” does not. Before any data is collected, agree on the decision the diagnosis will inform and who sees the raw data. A diagnosis with no named decision tends to grow into a survey of everything, and a diagnosis whose data goes only to senior leaders teaches the people who supplied it not to be candid next time.
Existing sources usually go further than a new survey: engagement results, exit interviews, incident logs, approval timelines and meeting calendars. Sorting them against a chosen model’s variables for two weeks often shows which variables have evidence and which are guesswork, and that gap is itself a finding about what the organization does not know about itself.
One boundary case deserves naming up front. When a sponsor has already decided the answer, a diagnosis becomes a justification exercise, and every model can be bent to support a conclusion reached in advance. Asking the sponsor at the start what finding would change their mind separates an open diagnosis from a ratification. If nothing would, a short conversation about that fact is worth more than any model.
The next step after the first pass is a choice that stays with the unit: act on one finding as a small experiment with a visible result, run a second lens on the finding nobody trusts, or stop because the problem turned out to be smaller than feared. Each of those is a legitimate outcome, and an organization that treats all three as success builds a habit of looking before it acts.
Anonymous. Counted, not tracked.
Where is your organisation with this right now?
What is the hardest part where you are?
In a sentence: what are you trying to work out right now?
No names, no company. Anonymous. Counted, not tracked.