Organizational Health Assessment for Change Planning
Learn how an organizational health assessment diagnoses performance and guides change planning, readiness, and strategy for lasting improvement.
Most leadership teams that commission an Organizational Health Assessment spend months choosing the survey and an afternoon deciding what to do with it. That allocation of effort runs the wrong way round. The index you license, the theory behind its items, the records you check it against and the people who interpret the report each change what the final reading can tell you, and most of those choices are made by default.
What an Organizational Health Assessment Measures That Performance Metrics Miss
An organizational health assessment estimates an organization’s capacity to keep performing in the future, whereas performance metrics such as revenue, margin and output report results the organization has already produced, so the two readings can point in different directions for years.
That gap is awkward for executive teams, because the reading they trust most is the one that arrives last. McKinsey defines organizational health as how well an organization aligns around a common direction, adapts to external shifts, executes with excellence and renews itself to sustain its performance aspirations, and its materials call health a leading indicator of the ability to achieve sustained improvements in performance. Four verbs, each describing something the organization does repeatedly, and none of them describing a number on a statement.
The academic literature arrives at a similar place from a different direction. The 2014 paper Organizational Health: Definition and Assessment, published in Procedia (Elsevier), set out to integrate existing concepts into a broader definition and to propose a metric of an organization’s capacity to perform and adapt to a dynamic internal and external environment. What the Procedia framing adds to McKinsey’s is the environment. Health is judged relative to conditions inside and outside the organization, which means a reading taken before a market shift or a reorganization can stop describing the organization the day after it.
Who Commissions a Health Assessment and When
Two sponsors commission most health assessments, and they arrive at the decision from opposite ends: executives after results that look too good to question, and HR leaders after a people signal that has stopped looking temporary.
The sponsor matters more than it seems. Whoever commissions the work tends to define its scope, and the scope decides which conditions the instrument can see at all. An executive sponsor tends to ask about direction, coordination and execution. An HR sponsor tends to ask about engagement, workload and retention. Both sets of questions belong in a health reading, and each sponsor on their own will usually leave half of them out.
Boards and Executives After a Strong Year
A board that has just approved a record set of results has the least reason to question the organization behind them, and that is why a strong year is a sound moment to look. The financial year is closed, budgets for the next one are being drawn up, and leaders have room to hear uncomfortable findings without the pressure of a crisis. The question an executive team brings is usually whether the conditions that produced the year are still in place: whether the people who carried the key accounts are staying, whether decisions still move quickly, whether the strategy is understood below the top two layers.
The risk is that a strong year produces a defensive reading of the results. Leaders who expect confirmation tend to explain away low scores as noise or as a vocal minority, because the financials seem to contradict them. Executives get the most from an assessment commissioned after a good year when they agree in advance which findings would change next year’s plan, such as a coordination score low enough to delay a planned expansion into a new region.
HR Leaders Facing Rising Attrition
HR leaders usually commission a health assessment after attrition, absence or exit-interview themes start to rise and the usual explanations stop holding. Turnover data is a count of decisions people have already made, and exit interviews hear only from those who have left. An assessment asks the people who are still there about the conditions that are pushing their colleagues out, which gives HR a chance to act before the next wave of resignations shows up in the workforce report.
The common trap for an HR sponsor is scope. An HR-led assessment framed around attrition tends to shrink into an engagement survey with a new name, measuring how people feel about their work without asking about direction, decision rights or coordination. When an HR team widens the scope to the management practices behind the attrition, the findings stop reading as an HR problem and start reading as an operating problem, and line leaders become far more likely to treat the results as theirs to act on.
Profitable but Unhealthy: When the Two Readings Diverge
A company can post strong profits while its health declines, because profit draws on assets built in earlier years (customer relationships, reputation, contracts, accumulated know-how) that keep paying out for a time after the conditions that built them have weakened.
AIHR’s guide to organizational health metrics describes organizational health as a broad umbrella term whose meaning shifts with business model and industry, and notes that a company could be very profitable yet still carry a toxic culture and a high turnover rate. The converse appears in the same guide: a company with low profits can have an excellent culture and very low turnover. Both cases describe the same mechanism running in different directions.
Consider how the divergence plays out. A sales organization keeps hitting its numbers because long-standing customers renew on contracts signed years ago. Meanwhile, its most experienced account managers are leaving, and the people replacing them have not yet built the relationships that drove the renewals. Revenue stays flat or grows for two or three years while the capability behind it thins. When the contracts come up for renewal under the new account managers, the decline arrives all at once, and the financial reports show it as a sudden event when it has been building for years.
Engagement and culture scores behave as leading indicators in this sequence: they move first, often a year or more before revenue does. A drop in how strongly people identify with the organization, or a rise in how many describe the culture as political, appears in survey data while the income statement still looks healthy. The difficulty is that leading indicators are easy to dismiss when the lagging ones look good, which is why the profitable-but-unhealthy pattern so often persists until the lag catches up.
Why Health Is Broader Than Culture or Engagement Alone
Culture and engagement are both parts of health, but neither covers how the organization sets direction, coordinates work, makes decisions or renews its capabilities, which is where many health problems actually sit.
An engagement survey measures an attitude: how committed and energized people feel. A culture survey measures shared norms: what behavior is rewarded, tolerated or punished. A health reading includes both and adds the operating conditions that engagement and culture scores cannot see. An organization can have highly engaged people working inside a decision process so tangled that their energy goes into waiting for approvals.
Bain’s Score Your Organization (Bain & Company) makes this concrete. The authors argue that decisions are the key to organizational performance and break decision effectiveness into distinct components, starting with quality (whether a company makes good decisions more often than not) and speed. Their example is Hospira, a specialty medical device and pharmaceutical company with a growth plan and a cost-savings target, whose marketing brochures moved through the organization in manila folders, with comments added in longhand and nobody clear on who had the final say. No engagement score would have caught that. People may well have been committed to the work; the problem lived in decision rights and process.
Organizational culture, then, is one input among several. Treating health as a synonym for culture leads leaders to commission culture programs when the binding constraint is a decision process, a coordination gap or a capability the organization has stopped building. The broader definition costs more to measure, and it is the only one that can locate a problem the organization did not already suspect.
Inside the McKinsey Organizational Health Index: Outcomes, Practices and Benchmarks
The McKinsey Organizational Health Index, the best-known third-party instrument for an organizational health assessment, is built in layers: nine health outcomes, the management practices that produce them, employee experience dimensions in the corporate version, and a benchmark database, followed by a headline claim about shareholder returns.
Each layer answers a different question, and leaders who quote only the headline tend to buy the instrument for the wrong reason. The outcomes say where the organization is strong or weak, the practices say which everyday behaviors explain that pattern, and the comparison database places the result against other organizations. McKinsey publishes the index in two configurations, one for companies and one for nonprofits, and the differences between them show which parts of the design are fixed and which are adjusted for the setting.
The Nine Outcomes and Where Each Comes From
The OHI scores nine outcome areas: direction, leadership, work environment, accountability, coordination and control, capabilities, motivation, innovation and learning, and external orientation, as listed on McKinsey’s Organizational Health Index program page.
The nine outcomes fall into four groups that track the align, execute and renew logic behind the index. Direction, leadership and accountability are about setting a course and holding people to it. Work environment, motivation and capabilities are about whether people have the climate, drive and skills to do the work. Coordination and control is about execution: whether performance is measured, risks are managed and work moves across units. Innovation and learning and external orientation are about renewal, the organization’s ability to change itself and stay connected to customers, partners and markets.
Reading the outcomes in these groups helps with interpretation. A profile that is strong on the first group and weak on the last tends to describe an organization that executes today’s strategy well and is slow to notice when that strategy stops working.
Direction, Leadership and Accountability
These three outcomes measure whether the organization knows where it is going, whether leaders shape the behavior needed to get there, and whether people understand what they are responsible for and face consequences tied to it. Direction covers whether a clear, shared view of the future exists and is understood below the top team. Leadership covers whether managers inspire and guide the actions the direction requires. Accountability covers whether roles, performance expectations and ownership are clear enough that individuals can be held to them.
The three depend on each other, which is why reading them together matters. High direction with low accountability tends to describe an organization where everyone can recite the strategy and nobody owns the parts of it that are slipping. Low direction with high accountability often produces the opposite: people deliver hard against objectives that do not add up to anything coherent. A leadership team that sees one of these patterns has a more specific conversation to hold than one that sees only an average across the three.
Innovation, Learning and External Orientation
Innovation and learning measures whether the organization generates new ideas and improves how it works, while external orientation measures how well it engages customers, suppliers, partners and other stakeholders outside its walls. Both outcomes face outward or forward. They capture the renewal half of health: whether the organization is changing fast enough to stay matched to its environment.
These are the outcomes most likely to be neglected in a strong year, because their payoff is slow and their absence is quiet. A company that scores low on external orientation may keep shipping products customers liked three years ago without noticing that preferences have moved. A low innovation and learning score paired with strong coordination and control scores is a common profile in mature, efficient organizations, and it tends to predict that the next disruption will be met with tighter controls when it calls for new capabilities.
43 Practices or 55: The Corporate and Nonprofit Configurations
McKinsey publishes two versions of the index that share the same nine outcomes but differ in the practice layer: 43 management practices in the corporate OHI, and 55 in the OHI for Nonprofits program.
McKinsey’s organizational health overview describes the corporate index as nine health outcomes, 43 management practices and 10 employee experience dimensions. The practices identify how on-the-ground behaviors affect health and help prioritize which behaviors to improve; the employee experience dimensions describe how employees experience their working lives. The nonprofit program measures the same nine outcomes against 55 management practices and compares each organization with a nonprofit benchmark, and McKinsey describes it as built on more than 20 years of research.
The nonprofit program page also sets out a timeline, which is the most concrete public account of what fielding the index involves.
| Feature | Corporate OHI | OHI for Nonprofits |
|---|---|---|
| Health outcomes | 9 | 9 (same list) |
| Management practices | 43 | 55 |
| Employee experience layer | 10 dimensions | Not listed on the program page |
| Comparison group | McKinsey client database | McKinsey nonprofit benchmark |
| Setup | Not published in the overview | 1-2 weeks of application and setup |
| Survey window | Not published in the overview | 1-day launch, then 2-3 weeks of fieldwork |
| Results | Not published in the overview | About 2 weeks to receive and explore results, then ongoing action prioritization |
The difference in practice counts tells a reader something useful. The outcomes are treated as universal, while the behaviors that produce them are adjusted to the setting: a nonprofit’s practices for funding, volunteers or mission focus have no exact corporate equivalent. A leadership team comparing its scores with published examples should check which configuration produced them, since a practice score from one version cannot be read against the other.
What the Benchmark Does to a Raw Score
What the OHI adds to a survey is the comparison: a raw score on accountability or motivation says little on its own, and the same score placed against a large reference population says a great deal.
The corporate overview states that the index compares an organization with a database of 2,600 clients and 8 million survey responses. That database is what turns a percentage of favorable answers into a position. Without it, a leadership team has no way to tell whether 62% agreement on a coordination item is a good result for an organization of its type or an alarming one. With it, the team can see whether it sits in the top quartile, the median or the bottom quartile on each outcome and practice.
The benchmark has its own limits. Its value depends on how closely the reference population resembles the organization being measured, so a public agency or a fast-growing startup may find itself compared with organizations whose conditions differ from its own. Benchmarks also pull attention toward relative position: leaders start asking how to move up a quartile and stop asking whether a particular practice matters for their strategy. And because the database is proprietary, the organization cannot inspect who is in it or how it has changed over time, so it has to take the composition of its comparison group on trust.
Reading the Three-Times Shareholder Return Claim
McKinsey reports that companies with top-quartile health deliver, on average, three times greater total shareholder returns over time, a figure whose comparison group, causal direction and time span all need checking before anyone quotes it.
The figure appears in the same How OHI works overview that describes the benchmark. As the vendor of the index, McKinsey is reporting a result about its own product’s predictive value. That does not make it wrong, and it does set the questions a careful reader should ask before quoting it.
The first question is about the comparison. AIHR’s metrics guide cites a version of the finding that compares healthy and unhealthy companies between 2003 and 2011, which is a different contrast from top quartile against everyone else. The second is about direction of cause. Companies with strong returns have more money and slack to invest in the practices the index rewards, so part of the association may run from returns to health. The third is about the phrase “over time”, which does not say over what period, or whether the return gap holds through downturns.
None of this removes the claim’s value as a reason to take health seriously. What it rules out is using the figure as a forecast for one organization’s share price. A leadership team that cites it in a board paper should pair it with its own evidence of what health has meant for its results.
Alternatives to the OHI: OHBI, COHA, TREC and nRhythm Compared
Organizations that cannot or will not license the McKinsey index can run an organizational health assessment on one of four other instruments: the peer-reviewed Organizational Health Behavior Index, NCWWI’s child-welfare workforce tool, TREC’s environmental-sector self-assessment, or nRhythm’s commercial scoring reports.
The four differ in who built them, what evidence stands behind them and what their output looks like, and those differences matter more than any comparison of which one is best. A public agency, a child welfare department, an environmental nonprofit and a mid-sized firm face different constraints on budget, sector fit and consultant support, and each instrument was designed with one of those situations in mind.
The OHBI’s Eight Dimensions and Its 2025 Validation
The Organizational Health Behavior Index measures eight dimensions: awareness, appreciation, relations, engagement, internal communication satisfaction, culture, persona and employee voice, split across a quantitative scale and a qualitative one.
The 2025 Frontiers in Psychology article Development and validation of the organizational health behavior index: a mixed-methods instrument for measuring organizational health, by Abad Alzuman, Lina Ibrahim Bakadam and colleagues at Princess Nourah Bint Abdulrahman University in Riyadh, sets out the instrument’s development and validation. Its quantitative subscale covers awareness, appreciation, relations, employee engagement and communication satisfaction through rating items. Its qualitative subscale covers organizational culture, employee persona and employee voice through open questions.
The index builds on an earlier version. In Sustainability (MDPI), the 2023 article Organizational Health Behavior Index (OHBI): A Tool for Measuring Organizational Health argued for assessing health with an index that combines qualitative and quantitative measurement, so that results help organizations respond to difficulties and opportunities in a changing environment. The 2025 study refined and validated that design.
The mixed design is the index’s main distinction. Most commercial health surveys rely on rating scales because they are easy to aggregate and benchmark. The OHBI keeps a qualitative half precisely for the dimensions that rating scales flatten: what kind of person the organization rewards, what culture people describe in their own words, and whether they believe their voice changes anything. The cost is analysis time. Open responses have to be read and coded, which an internal team must budget for, and an organization that skips that step ends up running half an instrument.
Sector-Built Tools
Sector-built tools trade breadth of comparison for fit: they are designed around the workforce and governance realities of one field, so their items speak the language of that field and their results point to interventions it already uses.
Two sector tools show the pattern. One comes from child welfare, a field with high caseloads, high turnover and a well-developed workforce research base. The other comes from environmental organizations, many of them small nonprofits with limited HR capacity. Neither offers a large cross-industry benchmark, and both are designed to be useful without one.
COHA for Child Welfare Agencies
The National Child Welfare Workforce Institute’s Comprehensive Organizational Health Assessment (COHA) helps child welfare agencies identify workforce strengths and gaps affecting the overall functioning of the organization. NCWWI presents it as a complement to its Workforce Development Framework, so the tool’s findings connect directly to the categories of intervention the framework already describes, such as recruitment, supervision and organizational climate.
Its most useful design feature is repeat use. NCWWI states that COHA can be used as a pre-post assessment to measure changes in organizational health resulting from targeted interventions. For an agency under pressure to show that a supervision program or caseload change made a difference, that turns the assessment from a one-off snapshot into an evaluation instrument: measure before the intervention, measure after, and compare the units that received it with those that did not.
TREC’s Organizational Health Review for Environmental Groups
TREC offers environmental organizations an Organizational Health Review, a self-assessment paired with its Org Ecology Model. A self-assessment of this kind is completed by the organization’s own staff and board without an outside consultant, which suits small environmental nonprofits that could never fund a proprietary survey. The ecology framing treats the organization as a living system whose parts (people, structures, relationships, resources) affect one another.
The self-assessment format carries a clear trade-off. It is cheap, fast and owned by the people who will act on it, and it gives up external comparison entirely. That makes it a good fit for an organization whose question is internal: where are we stretched, what needs attention before we grow. It is a poor fit for a funder or board that wants to know how the organization compares with peers, a question a self-assessment was never designed to answer.
Scores, States and Trends in nRhythm’s Reports
nRhythm’s Organizational Health Assessment reports category scores, an overall score and a trend score, all on a 0 to 5 scale that maps to three health states.
The category scores cover context, structures, work, energy, information and receptivity, which places nRhythm closer to an organizational design lens than to an employee survey: several of its categories describe how the organization is built rather than how people feel about it. The overall score summarizes the categories into a single position on the scale, and the three health states give that position a plain-language label.
The trend score is the distinctive element. nRhythm says it is meant to predict future states and guide proactive redesign, which makes it a different kind of reading from the overall score. An overall score describes where the organization sits now. A trend score describes which way it is moving. Two organizations with the same overall score can have opposite trends: one recovering from a difficult reorganization, the other sliding from a strong position. For a leadership team deciding whether to act now or watch, the direction of travel often matters more than the level.
The main limit of the reports is transparency. nRhythm’s public material does not explain how the overall score is derived from the categories or where the thresholds between health states sit, so a buyer has to ask how a 3.2 differs from a 3.4 and whether the gap means anything before using either number to make a decision.
Matching the Instrument to the Decision
Choosing among these instruments depends on whether the organization needs an external benchmark, a sector-specific workforce lens, or a self-assessment it can run without a consultant, so the starting point is the decision the results must inform.
| Instrument | Built for | Design | External comparison | Fits the decision |
|---|---|---|---|---|
| McKinsey OHI | Companies and nonprofits licensing a proprietary index | Outcomes, practices and benchmark | Large proprietary database | Where do we stand against many other organizations? |
| OHBI | Researchers and organizations wanting a validated public instrument | Mixed quantitative and qualitative subscales | None built in | What behaviors shape our health, in people’s own words? |
| COHA | Child welfare agencies | Workforce assessment, repeatable pre-post | Within the agency over time | Did our workforce intervention change anything? |
| TREC Organizational Health Review | Environmental organizations | Self-assessment with the Org Ecology Model | None | Where are we stretched before we grow? |
| nRhythm | Firms seeking a design-oriented commercial report | Category, overall and trend scores on a 0 to 5 scale | Not described publicly | Which way are we moving, and what should we redesign? |
Read across the rows, the instruments sort less by quality than by purpose. An agency evaluating an intervention gains nothing from a cross-industry benchmark, and a board asking for peer comparison gains little from a self-assessment. Mismatches are common because organizations tend to choose the instrument they have heard of first, then discover after fieldwork that it answers a question nobody asked.
The Theories Underneath Health Instruments: JD-R, Social Exchange and Competing Values
Most organizational health assessment instruments draw their items from a small set of theories, and three of them, the Job Demands-Resources model, social exchange theory and the Competing Values Framework, explain much of what a modern health survey asks and what a low score is actually claiming.
The OHBI makes the link explicit. Its 2025 validation study, whose authors include Muath Al Jaafari, names all three as foundations: the JD-R model for linking organizational resources to engagement and well-being, social exchange theory for the reciprocal relationship between organization and employee, and the Competing Values Framework for placing culture among competing orientations. That combination is why the index’s items span resources, relationships and culture at once. Knowing the theory behind an item tells a leader what the item assumes, and assumptions are where interpretation goes wrong.
Demands, Resources and the Strain Between Them
The Job Demands-Resources model reads health as a balance: demands such as workload, ambiguity and emotional strain on one side, and resources such as autonomy, support, feedback and learning on the other.
The OHBI authors trace the model to Arnold Bakker and Evangelia Demerouti and describe it as the foundation for linking resources such as awareness, appreciation and communication to engagement and well-being, with resources buffering demands and fostering motivation. In survey terms, every item about workload or role clarity is a demands item, and every item about support, development or discretion is a resources item. A health reading built on JD-R looks at the ratio between them.
When a Fix for Workload Removes Autonomy
Demands and resources are context-dependent, so a process introduced to reduce workload can remove autonomy or learning and create new strain. A team overwhelmed by ad hoc requests gets a ticketing system and a standard intake form. The volume of interruptions drops, which looks like a reduction in demands. But the same system takes away the team’s discretion over which requests to take on and removes the conversations in which junior members learned how senior colleagues judged priorities.
A complexity reading of JD-R asks what each intervention enables and what it constrains, because the answer is rarely all one or the other. The practical consequence for an assessment is that a falling workload score paired with a falling autonomy or development score should be read as a trade-off the organization made, possibly without noticing. Treating each score as a separate problem tends to produce a second fix that undoes the first.
Resources That Buffer Emotional Strain
Some demands cannot be reduced: a child protection worker’s caseload carries emotional strain however well it is managed, as does an emergency nurse’s shift. In those settings JD-R points to resources that buffer the strain: supervision that gives workers a place to process difficult cases, peer support, and enough autonomy to control how the work is done even when its content is fixed.
This changes what a low well-being score means in emotionally demanding work. It rarely signals that demands are too high in a way leaders can simply lower; it more often signals that buffering resources are thin or have eroded. An assessment that reads emotional strain items alongside supervision and peer-support items can tell the difference between work that is hard and work that has become unsupported, and only the second points to something the organization can change quickly.
Reciprocity: Why Appreciation and Voice Appear in Surveys
Social exchange theory explains why appreciation, relationships and voice appear in health instruments: employees reciprocate the support they perceive from the organization with effort and commitment, and withdraw both when the exchange feels one-sided.
The OHBI authors draw on Blau’s formulation of the theory to explain the reciprocal relationship between organizations and employees, in which supportive behaviors such as recognition and appreciation lead to greater commitment and trust. Each survey item on appreciation asks, in effect, whether the employee believes the organization is keeping its side of the exchange. Each item on relationships asks whether the exchange holds with immediate colleagues and managers, where most of it actually happens.
Voice items test a particular part of the exchange. When someone raises a concern or an idea, they are offering the organization something: information, effort, a risk to their own standing. A voice item asks whether that offer is received. People who conclude that it is not stop making the offer, and the organization loses information it does not know it is missing.
The theory explains why low appreciation and voice scores matter beyond morale. Reciprocity predicts that when perceived support drops, discretionary effort goes first: the extra hour, the problem flagged early, the colleague helped without being asked. That withdrawal is invisible in output data for months, because people still do the defined parts of their jobs. A health assessment that reads appreciation and voice scores through social exchange theory can see the exchange weakening while performance figures still look normal, and that gives leaders time to act on the support side before the effort side responds.
Holding Tensions Instead of Choosing One Culture
The Competing Values Framework treats culture as a set of tensions among collaboration, control, competition and creativity, so a healthy culture is one that can hold those values in productive balance instead of maximizing any one of them.
The OHBI study cites the framework through Kim Cameron and Robert Quinn’s work, placing culture within four competing orientations: clan, adhocracy, market and hierarchy. Because the four sit in tension, an organization rarely belongs to one type; it carries a mix, and the mix shifts as leaders push one orientation harder than the rest. The framework’s design point is balance among the four. Each orientation is useful, and each becomes a problem when it crowds out the others.
Read as a health question, the framework asks which value dominates, what it makes possible, and what it makes invisible. A hierarchy-dominant organization makes reliability and compliance possible and makes experimentation invisible: nobody proposes a new approach because no process exists for it. A market-dominant organization makes speed and results possible and can make collaboration invisible, since helping another unit earns no credit. A clan-dominant organization makes loyalty and support possible and can make underperformance invisible, because nobody wants to name it.
This reframing changes what an assessment reports. A typing exercise that labels the organization as having a clan culture tells leaders very little. A health reading that shows which orientation dominates and which ones it is suppressing tells leaders where the organization’s blind spots are, and blind spots are where the next serious problem tends to start.
Pairing Survey Scores With Hard Data: Sickness Absence and Workplace Health Indicators
An organizational health assessment that pairs survey scores with hard records such as sickness absence, workplace health indicators and safety climate measures can test whether what employees report is reaching their health and behavior, because a survey captures perception while absence, turnover and safety records show what that perception becomes.
Recent research gives three working models for the pairing: a Swedish public-sector study that set survey responses beside sickness-absence records, a set of workplace health indicators built for enterprise assessment, and a body of work on safety climate and health culture tools. Each model also raises the practical question of what to do when the two kinds of evidence do not agree.
The Umeå Public-Sector Design: Surveys Plus Sickness Absence
Researchers led from Umeå University combined employee survey items with municipal sickness-absence records to test whether a survey-based definition of a healthy workplace matched what the absence data showed.
The 2026 article Organizational health in public-sector workplaces: a mixed-method study, published in Frontiers in Organizational Psychology by Elin Forsell, Magnus Akerstrom, Susanne Tafvelin, Andreas Tornevi and Jens Wahlström, used three kinds of data from one Swedish municipality: focus groups with stakeholders and work environment specialists, survey items on self-assessed production loss due to health and work environment problems, and sickness-absence statistics.
Their design uses sickness absence as an outcome indicator that can confirm or challenge what employees report. The team defined healthy workplaces from the survey responses, then checked whether those workplaces had lower sickness absence than the rest. They did, with the clearer difference in long-term absence, which suggests the survey was picking up conditions that eventually show up in people’s health. The focus groups added a further finding: organizational health appears at several levels at once, at the organizational level as sufficient preconditions for meeting objectives, at the workplace level as team capability and stability, and at the individual level.
A mixed-methods assessment of this kind is replicable by most public employers, since sickness-absence records already exist and survey items on production loss are short. The authors note its limits, including that sickness absence is an imperfect measure and that the findings come from a single municipality. For an internal team, the lesson is the structure more than the result: decide in advance which record would confirm the survey, then check whether it does.
Indicators Beyond the Survey
Beyond sickness absence, an assessment can draw on indicator sets designed for enterprise health measurement and on tools that measure safety climate and health culture, both of which extend the evidence past what employees say in a questionnaire.
Deloitte makes the starting point practical. Its article The importance of proactive organizational health assessments (Deloitte) notes that a high-level understanding can be obtained by analyzing existing organization data, and that using existing data and analysis tools, organizational health can be diagnosed in just a few weeks. Spans of control, management layers, vacancy durations and internal mobility rates are all records most organizations already hold. Read together, they show structural conditions that a survey cannot.
Sustainable Workplace Health Indicators
Recent occupational health research has proposed sets of sustainable workplace health indicators for enterprise assessment. An indicator set of this kind defines in advance which measurable conditions count as evidence of a healthy workplace, so that assessment rests on a stable list of measures instead of whatever data happens to be available that year. The word sustainable signals the intent: indicators that can be tracked across years and compared between sites.
For an organization running repeated assessments, a defined indicator set solves a common problem. Without one, each round tends to add and drop measures as sponsors change, and trend lines break. With one, the assessment can show whether a site’s health indicators moved after an intervention, and a multi-site employer can see which locations diverge from the rest. The practical step is to agree on the indicator list before the first round and protect it from revision for at least two cycles.
Safety Climate and Health Culture Tools
Fan-Yun Lan and Justin Yang edited a 2026 research collection on evaluating organizational health culture, introduced in their editorial Evaluating organizational health culture: tools and impact on workplace health interventions in Frontiers in Environmental Health. The collection examines how organizational health culture, safety climate and structural integration affect the success of workplace health interventions across industries such as healthcare, construction and emergency services. Safety climate measures capture how employees perceive the priority their organization actually gives to safety, as distinct from what its policies say.
The editorial concludes that effective workplace health interventions require culturally integrated environments. For a health assessment, that makes safety climate a useful bridge measure: it is a perception, like other survey items, but it is tightly linked to a hard record (incident and near-miss rates) that can test it. An organization whose safety climate scores are high while near-miss reporting falls has a question to investigate, and a health program launched into a weak health culture is likely to be ignored however well it is designed.
When Perception and Records Disagree
When survey scores and hard records disagree, the disagreement is a finding in its own right: it points to a condition that neither source shows clearly on its own, and the work is to work out which condition explains the gap.
Linking survey responses to recorded data unit by unit, as the Umeå team did, is what makes a disagreement visible in the first place, since an organization-wide average of each source can hide opposite movements in different units. The common patterns, and the explanations each one usually calls for, are these:
| Survey reading | Records reading | Likely explanations to test |
|---|---|---|
| Positive | Absence or turnover rising | People fear candid answers; presenteeism hides strain until it becomes long-term absence; the survey missed a group |
| Negative | Absence and turnover stable | Strain has not yet turned into illness; people are staying for external reasons such as a weak job market; absence is under-recorded |
| Positive | Safety incidents rising | Safety climate is reported as it should be, not as it is; incident reporting improved |
| Negative | Records improving | A recent intervention is working faster in behavior than in perception; survey timing caught a bad week |
Every row in the table is a hypothesis to test. Each row has more than one explanation, and the way to choose between them is to take the discrepancy to the people in the unit concerned and ask. A team whose survey is positive while absence rises can usually say within one conversation whether the survey felt safe to answer or whether a few long-term cases are skewing the record. Skipping that step and averaging the two readings into one figure loses the very information the pairing was designed to produce.
How to Run an Organizational Health Assessment From Scoping to Action Planning
Running an organizational health assessment with internal resources takes a sequence of stages, from purpose and scope through instrument choice, fieldwork, analysis and feedback to action planning, and AIHR’s 11-step framing of organizational assessment shows how small a share of the work the survey itself is.
AIHR’s guide Organizational Assessment: Types and 11 Steps For Conducting Yours describes organizational assessment as a comprehensive audit of an organization’s people, processes and platforms, and lays the work out across eleven steps. Choosing and fielding an instrument occupies only part of that sequence. For an internal HR or OD team, the stages of the organizational health assessment process group as follows:
- Purpose and scope: the decision the assessment informs, who is included, and who owns the results.
- Instrument choice: matched to that decision, as compared earlier.
- Fieldwork: survey, interviews, records, and a map of the informal organization.
- Analysis: patterns across outcomes, units and data sources, and tests of fit between the formal design and how work actually flows.
- Feedback and action planning: results returned to the people who supplied them, a short list of priorities, and a scheduled repeat measure.
Scoping: Purpose, Population and Ownership of Results
Scoping fixes three things before any instrument is chosen: the decision the assessment will inform, the population it will cover, and who will see and own the results.
Purpose comes first, and it should be written as a decision instead of a topic. “Understand our culture” is a topic; “decide whether to restructure the regional operations or invest in their management capability” is a decision. Writing the purpose as a decision forces the sponsor to say what they will do with the answer. It also tells the team which outcomes and practices the instrument must cover and which it can leave out.
Population is the second scoping choice. A census of the whole organization gives every unit its own results; a sample is cheaper but may leave smaller units without readable data. The binding constraint is usually anonymity: any reporting group must be large enough that nobody can be identified from the results, which in practice means small teams are combined before reporting. Deciding those combinations in advance prevents arguments after the data arrives.
Ownership is where many assessments go wrong without anyone noticing. Accountability is essential to every health assessment, which means telling the whole organization what is being measured, who will see which results, and how the data will and will not be used, before the first question is asked. It also means naming which leader owns the response for each part of the organization. An assessment that leaves ownership vague tends to produce a report that every leader reads and none acts on, because each assumes the issues belong to someone else.
Seeing the Informal Organization
A survey measures how individuals experience the formal organization, but much of how work actually gets done runs through informal channels (who people ask for advice, whom they trust, which routes information really takes) that no questionnaire item captures directly.
Two diagnostic lenses bring the informal organization into an assessment. One maps the relationships themselves. The other tests whether the formal design and the informal reality pull in the same direction. Used together, they explain survey results that otherwise look contradictory, such as a unit with high scores on leadership and low scores on coordination.
Network Analysis of Advice and Trust
Organizational network analysis maps advice, trust, collaboration and information flow by asking people short questions about whom they go to for help, whom they trust with a problem, and whom they share information with. The answers form a network, and the network reveals brokers who connect otherwise separate groups, isolates who are cut off from the flow, and overloaded connectors through whom too much work and information passes. None of these roles appears on a formal org chart.
The findings often explain survey results directly. A unit with low coordination scores may turn out to depend on a single broker who links it to the rest of the organization, so every cross-unit request waits on one person. An overloaded connector may be the reason a capable team reports high workload while its formal headcount looks adequate. Network analysis gives the assessment a specific intervention point (redistribute the broker’s load, create a second link) where the survey alone would offer only a general coordination problem.
Congruence Between Formal and Informal Systems
The Nadler-Tushman Congruence Model tests fit among four components: the work, the people, the formal organization and the informal organization. The Change Compass guide How to Diagnose for Change describes it as examining the fit between work, people and formal and informal organization, and recommends it when performance problems seem to come from mismatches. The model’s core claim is that a sound strategy can still stall where incentives and norms pull against it.
The McKinsey 7-S Framework adds breadth across strategy, structure, systems, shared values, skills, style and staff; the same Change Compass guide recommends it for testing whether a restructure or new strategy is matched by the softer elements. The framework works best when its elements are read as mutually adjusting variables, since changing structure shifts style and skills whether or not anyone plans it. In an assessment, both models turn scattered scores into a question about fit: a new collaborative strategy with individual bonus incentives shows up as low coordination scores, and the fix lies in the incentive system rather than in a coordination workshop.
Feedback, Action Planning and the Repeat Measure
The assessment’s final stage returns results to the people who supplied them, commits to a few priority practices, and schedules a repeat measure to test whether those practices moved.
Feedback has to reach every group that took part, in a form they can read and discuss, within weeks of the survey closing. Results that disappear into an executive pack teach employees that their answers had no effect, and they respond to the next survey with lower participation and safer answers. Each round of an assessment depends on the credibility left by the one before it.
Action planning should commit to a small number of priority practices, usually two or three across the organization and one or two per unit. A long list of actions spreads effort thinly, and none of the actions shifts far enough to register in the next measurement. The priorities should come from the patterns in the data and the decision written at scoping, and each should name an owner, an observable change in behavior, and a date.
The repeat measure closes the loop. Scheduling it at the same time as the action plan, typically twelve to eighteen months later for a full assessment, or sooner with a short pulse on the priority practices alone, turns the plan into a test. If the targeted practices moved and the related outcomes followed, the organization has learned something about its own system. If they did not move, it has learned something too, and the next plan can adjust.
Health Assessment or Change Diagnosis: Which Question Are You Asking?
A health assessment asks whether the organization has the general capacity to align, execute and renew, while a change diagnosis asks how a specific change will land on the groups it affects, and choosing the wrong one wastes the effort of both.
The confusion is common because the two use similar tools: surveys, interviews, readiness ratings. The difference lies in the unit of analysis. A health assessment’s subject is the organization as a whole, measured against a stable set of outcomes. A change diagnosis’s subject is one change and the people it touches, measured against what that change will demand of them. Leaders about to launch a program need to know which question the moment calls for, how to classify the situation before choosing, and how an existing health reading can shorten the diagnosis.
General Capacity Versus the Landing of One Change
A change diagnosis looks forward to one specific change and the groups it will hit, so its questions concern load, readiness and history that a general health reading does not ask about directly.
The Change Compass guide How to Diagnose for Change describes diagnosing for change as assessing the organization before the program begins: how much change the affected groups are already carrying, how aligned and ready leaders are, what past change has taught people to expect, and where the organization’s structures, systems and culture will help or block the new way of working. The guide adds that a sound diagnosis can still fail if the affected teams are absorbing three other changes in the same quarter. None of those questions has an answer in a general health score, because each depends on which groups the change touches.
The guide also borrows a clinical analogy in which organizational health resembles human wellness: when imbalances build up in the system, the organization becomes less able to absorb a change, much as a person run down by illness copes less well with a new demand. The analogy is useful for the sequence it implies. Medicine separates the patient’s general condition from the specific procedure, and checks the first before scheduling the second.
Project-level tools sit on the change side of the line. Prosci’s guide Using the PCT Model to Diagnose, Treat and Monitor Project Health describes the PCT Assessment’s 40 diagnostic questions, which gauge the health of one initiative across success, leadership and sponsorship, project management and change management. That is a health reading of a single project, and it answers a narrower question than an organizational health assessment does.
Classifying the Situation First
Before choosing either assessment, a leader can classify the kind of situation the organization faces, because the classification decides whether a structured diagnosis will help at all or whether the situation calls for probing and learning.
Poor diagnosis is a widely recognized weakness. HBS Online’s guide Root Cause Analysis: What It Is & How to Perform One (Harvard Business School Online) cites research described in Harvard Business Review finding that 85 percent of executives believe their organizations are bad at diagnosing problems and 87 percent think that flaw carries significant costs. Much of that weakness comes from applying one diagnostic approach to every situation. Two frameworks help a leader classify first and structure the inquiry second.
Cynefin’s Five Domains as a Pre-Assessment Filter
Dave Snowden’s Cynefin framework, presented by The Cynefin Co, distinguishes five kinds of situation: clear, where cause and effect are obvious and best practice applies; complicated, where cause and effect can be worked out by analysis and expertise; complex, where cause and effect are visible only in hindsight and the way forward is to probe with safe experiments; chaotic, where no stable pattern exists and the first need is to act to restore order; and confused, where it is not yet clear which of the other four applies.
As a filter before assessment, the framework changes what an assessment can be expected to deliver. In a complicated situation, a structured health assessment or change diagnosis can identify the problem and point to a fix. In a complex situation, the same assessment can only surface patterns and candidate experiments, because no analysis will reveal the answer in advance. A leader who classifies the situation as complex and still expects the assessment to produce a definitive plan will be disappointed by results that were always going to be exploratory.
EPIC: Explore, Produce, Intervene, Check
Critical Systems Practice structures inquiry through an EPIC sequence: Explore the situation from several perspectives, Produce an intervention strategy that combines suitable methods, Intervene flexibly, and Check whether the intervention achieved what was intended. The sequence turns assessment into a collective inquiry into boundaries (who and what is included), assumptions (what each perspective takes for granted) and power (whose view shapes the definition of the problem).
For a leader choosing between a health assessment and a change diagnosis, the Explore stage is where the choice gets made well. Exploring from several perspectives often reveals that the presenting problem, such as resistance to a new system, is a symptom of a wider health condition like low trust in leadership, or the reverse. The Check stage builds in what many assessments omit: an explicit return to whether the intervention worked, which turns a one-off diagnosis into a learning cycle.
Using a Health Baseline to Speed a Change Diagnosis
A recent health baseline speeds up a later change diagnosis because it already shows where capacity is thin before the specific change adds its own load, so the diagnosis can concentrate on the groups and questions the baseline flags.
The connection runs through specific outcomes. A low coordination and control score warns that a change requiring new cross-unit handoffs will struggle. A low leadership score in a particular division warns that sponsors there will need support before they can lead others through the change. A low innovation and learning score warns that people may lack the habits a change in ways of working depends on. Each of these narrows the change diagnosis to the places where it matters most.
The age of a baseline limits how far it can be trusted. A reading taken before a reorganization, an acquisition or a leadership turnover may describe an organization that no longer exists. As a working rule, a baseline older than about a year, or older than the last major structural change, should be checked with a short pulse before it shapes a change plan.
The baseline also cannot answer the change-specific questions on its own. It says nothing about how much other change is landing on the same groups, or what those groups learned from the last program. A leader who treats the baseline as a substitute for the diagnosis will plan around general capacity and be surprised by load and history that only the diagnosis would have found.
Reading Results as a Conversation Map, Not a League Table
Reading organizational health assessment results well rests on three moves: seeing health as several interacting conditions, widening the search for causes beyond motivation and leadership, and handing interpretation of each unit’s results back to the people who work there.
Each move answers a temptation that arrives with the finished report. The first is to compress the results into a single number. The second is to look for causes only in the outcomes the survey named. The third is to rank units against one another and treat the ranking as the finding.
Several Conditions, Not One Number
Complexity thinking treats health as several interacting conditions, because the dimensions an instrument measures affect one another, and the interactions often matter more than the levels.
An organization can have high engagement but poor voice: people care about the work and have stopped raising concerns because concerns go nowhere. It can have strong relationships but weak communication: colleagues trust one another while information still fails to cross unit boundaries. It can have a compelling culture that suppresses dissent, where the shared sense of mission makes disagreement feel like disloyalty. In each case, two scores that would average to a respectable figure describe a specific dynamic, and the dynamic is what a leader needs to see.
A composite score averages away exactly the contradictions that explain what is happening. High engagement and low voice combine into a middling figure that suggests nothing much is wrong, when the pairing describes a committed workforce withholding information. Organizations are complex adaptive systems, in which local conditions interact to produce patterns that no single dimension predicts, so the most useful reading of a health report is often found in the pairs of scores that disagree.
The practical consequence is a change in how results are presented. A report that leads with an overall health index invites ranking and comparison. A report that leads with the two or three most striking contradictions in each unit’s profile invites the question of what is producing them, which is the question an organizational health assessment exists to raise.
Widening Where to Look for Causes
When a score is low, the instinct is to look for the cause in the outcome itself: low motivation means a motivation problem, low leadership scores mean a leadership problem. Several frameworks widen the search to conditions that sit outside those obvious categories.
The NASSS framework and its Complexity Assessment Tool, developed in health services research, assess complexity across domains such as the condition being addressed, the technology involved and the value proposition. Their logic carries over to health results: a struggling initiative may be struggling because of the technology or the value it offers, not because of the people asked to adopt it. Two further lenses do the same work for health readings in general.
Learning and Knowledge Flow in the HCOL Compass
The HCOL Compass links three areas: leadership that reinforces learning, organizational context, and knowledge management and mobilization. Taken together, they recast the central question of health as whether the system can notice, share, interpret and act on local knowledge. A unit that scores well on capabilities but poorly on innovation and learning may have skilled people whose knowledge never travels, because no route exists for it to reach the people who could use it.
This lens shifts the search for causes from individuals to flows. When frontline staff see a problem early and leaders hear about it late, the cause is rarely a lack of skill or motivation on either side; it is usually a missing channel, a context that discourages raising issues, or leaders who do not ask. A health reading viewed through the HCOL Compass points toward interventions such as after-action reviews, cross-unit forums or changes to what leaders ask in team meetings, all aimed at the flow of knowledge instead of the people at either end.
Socio-Technical Friction in Employee Experience
Employee-experience measures span physical, technological, cultural, reputation and recruitment dimensions, so friction may start in tools, spaces or policies rather than in motivation or leadership. A team reporting low motivation may be fighting a slow internal system every day, working in an office designed for a different kind of work, or living with a policy that makes simple requests take weeks. These are socio-technical frictions: the interaction between people and the technical and physical systems they work within.
Reading health results alongside employee-experience data protects against a costly misdiagnosis. Leaders who see low motivation scores tend to respond with leadership training or engagement initiatives, and when the real cause is a broken expense system or an unworkable shared drive, those responses fail and add cynicism. Checking the technological and physical dimensions first often reveals a fix that is cheaper, faster and more visibly responsive than any program aimed at attitudes.
Handing Interpretation Back to the Teams
Reading results well means helping people recognize system patterns, identify leverage and choose the next test themselves, so each team interprets its own results before anyone above it draws conclusions.
The people inside a unit can explain its results in ways no analyst can. They know that the low coordination score reflects a handoff with one other team, that the drop in appreciation followed a specific reorganization, or that the high workload score comes from one recurring report. A facilitated session in which a team reads its own profile, names the patterns it recognizes and chooses one small experiment to run produces a specific, owned action instead of a generic recommendation from above.
The experiment should be small enough to try within a few weeks and observable enough that the team will know whether it worked: a change to how a weekly meeting runs, a new handoff agreement with a neighboring team, a trial of a different approval route. The team chooses it, runs it and reports what happened, and the next measurement shows whether the related scores moved.
Ranking teams against one another on a single composite number works against all of this. A league table turns the next survey into a contest that unit managers want to win, and the pressure reaches their teams as a signal about what the right answers are. Within a round or two, the scores start to reflect the ranking more than the work, and the assessment loses the honest signal it depended on.
How to Start Applying Organizational Health Assessment
For most organizations, the lowest-risk first step is a pilot in two or three units, using a free or low-cost instrument such as the OHBI or a sector tool where one exists. A pilot tests the process before the instrument: how long fieldwork actually takes, whether response rates hold, whether reporting groups are large enough to stay anonymous, and whether managers can lead a team readout without the session turning defensive. Those operational lessons are expensive to learn across a whole organization and cheap to learn in a pilot, and they decide whether a later full assessment, licensed or not, produces results anyone uses.
Timing deserves a deliberate choice of its own. An assessment fielded in the weeks after a layoff announcement, during an acquisition, or in the middle of annual performance reviews measures the shock of the moment more than the organization’s condition. If one of those events is unavoidable, the sensible options are to wait until it has settled, usually a quarter or more, or to field the assessment anyway and label the results explicitly as a reading taken under those conditions.
Survey fatigue is the failure mode that catches organizations already running engagement pulses, wellbeing checks and culture surveys. Adding a health assessment on top asks the same people similar questions under a different name. Before launching, list every employee survey run in the past year, retire or fold in those that overlap, and tell employees which ones have stopped. Fewer, better-connected surveys usually produce higher participation than a larger set that each claim to be essential.
Very small organizations face a boundary case. Below a few dozen people, most units are too small to report separately without losing anonymity, and a benchmarked index adds little. In that situation, a facilitated self-assessment with the whole team in the room, using an instrument’s items as conversation prompts instead of a scored survey, often yields more than a questionnaire could, because the people closest to the work can say directly what they see.
Finally, budget for the time after fieldwork. Internal teams routinely plan the survey in detail and leave analysis, team readouts and follow-up to whatever capacity remains. A rough rule is to allow at least as much staff time for interpretation and action as for design and fieldwork combined. Where that time cannot be found, a narrower assessment that the organization can fully act on is the better choice.
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.