Organizational Health: A Small Business Assessment

The first company I ran had good people, a growing customer list, and a problem I could not name for about a year. Decisions took three weeks. The same argument about pricing came back four separate times and was settled differently each time. Then two competent people resigned in the same month, and both said a version of the same sentence: nobody here knows who actually decides things.

I had been measuring the wrong thing. We ran a satisfaction check every quarter and it came back fine, because people liked each other and liked the work. What was broken was not how anyone felt. It was the organization itself: how it agreed on a direction, how it made a call, and how it corrected course when the call turned out to be wrong.

That is what organizational health measures, and it is the diagnostic small companies skip, because the instruments for it were built for organizations a hundred times their size. This guide covers what the term means, how it differs from engagement and culture, the six dimensions worth scoring, a 24-statement assessment you can run in a week, how to read the result, and which parts of the enterprise version are worth borrowing. I build the onboarding, records, and org chart layer that a lot of this rests on at FirstHR.

What Organizational Health Actually Means

Organizational health is how well a company can agree on what matters, deliver against that agreement, and change it when reality moves. It is a property of the system rather than of the people inside it, which is why a team of good performers can produce a badly run company and be genuinely puzzled about how.

The term entered general business use through Patrick Lencioni's book The Advantage, which argues that a healthy organization is one with minimal politics and confusion, high morale, and low unwanted turnover, and that this is a larger untapped advantage than strategy or technology. His four disciplines are building a cohesive leadership team, creating clarity, overcommunicating that clarity, and reinforcing it through the systems people actually touch.

The enterprise version of the same idea, sold as an organizational health index, organizes it into three attributes: internal alignment, quality of execution, and capacity for renewal. That framing holds up well. Its problem at small scale is granularity, since a company of eighteen people cannot act on a score for something as broad as alignment. The six dimensions further down are that same structure cut into pieces a founder can change on a Tuesday.

The most useful thing about the concept is what it rules out. Health is not morale, not perks, and not whether the last all-hands went well. It is closer to whether two people who depend on each other could describe this quarter's priority the same way, and whether the work one hands the other arrives ready to start.

Organizational Health vs Engagement, Culture, and Satisfaction

Engagement measures how invested people feel, culture describes the behavior a group repeats, satisfaction measures contentment with conditions, and organizational health measures whether the company can align, decide, and deliver. They correlate often enough to be confused with each other, and they fail in different directions.

The distinction is not academic, because the two results point at different work. A low engagement result sends you toward managers, recognition, and growth paths. A low health result sends you toward decision rights, priorities, and handoffs, which is a different set of fixes owned by a different person, usually you.

The combination worth watching is high engagement with low health, and it is more common at small companies than the reverse. People care, work hard, and still lose two days a week to unclear ownership and redone work. Engagement measurement cannot see that, because everybody involved answers the questions honestly and positively while the system quietly wastes their effort.

Why Small Teams Feel Poor Health First

At small headcount there is no slack to absorb a system problem, so poor organizational health converts into missed dates and departures within weeks rather than quarters. A large company can carry a broken handoff for a year inside a function nobody audits. A team of fifteen cannot carry it for a month.

Three things make the effect sharper. Every person is a larger share of total capacity, so one point of failure stops actual output. Roles overlap, which means ownership is genuinely ambiguous rather than merely undocumented. And the founder is usually a participant in every broken loop, which makes the problem hardest to see from the one seat that could fix it.

There is also a cost figure that reframes what a bad reading is worth. SHRM research on toxic workplace culture put the price of culture-driven turnover at $223 billion over five years, with 76 percent of workers saying their manager sets the culture and 36 percent saying that manager does not know how to lead a team. At a company where the manager is also the owner, both halves of that sentence describe the same person.

The manager point is worth stating plainly, because it decides who does the work of fixing a low score. Gallup finds that managers account for 70 percent of the variance in team engagement. In a business of twenty people, that manager is usually the founder, which makes an organizational health assessment less a survey of the team than a mirror pointed at how you run the place.

The Six Dimensions Worth Scoring

Six dimensions cover what actually breaks in a small organization: direction, decision rights, execution, coordination, capability, and renewal. They are the enterprise attributes of alignment, execution, and renewal cut into pieces small enough that each one has an obvious owner and an obvious first fix.

Two of the six do most of the damage at small scale. Direction fails first, because clarity feels complete from inside the founder's head and arrives at the team as three different versions of the same priority. Decision rights fails second, and it fails as the company grows: a structure that worked at eight people routes every call through one person at eighteen, and the queue becomes the constraint on everything else.

Capability is the dimension people most often misread as a hiring problem. A low score there usually means one person is the single point of failure for something the business cannot pause, which is a design problem rather than a talent problem. Writing down who covers what, and making the reporting structure explicit, resolves more of it than a new hire would.

Renewal is last on the list and first to be neglected, because its failure mode is silence. Nothing looks wrong in a company where bad news does not travel, right up until the resignation arrives with a full explanation attached. If you want one early signal, watch whether anything visible has changed in the past quarter because a person here spoke up.

The 24-Statement Assessment

Score four statements per dimension on a scale of 1 (strongly disagree) to 5 (strongly agree), which produces 4 to 20 points per dimension and 24 to 120 overall. It takes about ten minutes to answer, everyone answers it including you, and the wording stays identical between rounds so the comparison means something.

The statements are written in the first person and describe observable conditions rather than opinions about the company. That is deliberate. Asking whether we are aligned invites a diplomatic answer, while asking whether you could write down the top three priorities and have colleagues write the same three invites a real one.

The first sheet holds the 24 statements with columns for your score, the team average, and the gap between them. The second rolls the statements into the six dimension scores, with room for the lowest single statement and what the comments said. The third is the round history, which is the sheet that turns a one-off exercise into a trend and records what you changed after each round.

Add one open text question at the end and keep it constant too: what is the single thing that most slows your work down here? It is the only question in the set that can tell you about something you failed to ask about, and it consistently produces the most specific material in the whole exercise.

How to Run It Without Making Things Worse

Run it in a week: score it yourself first, send it anonymously with a note that says when people will hear back, close it after one reminder, and report the six dimension scores to the team within two weeks. The sequence matters more than the instrument, because most of the value comes from the comparison between your view and theirs.

One judgment call comes up every time: whether to break the results down by team. Below about twenty-five people, do not. The subgroups are too small to be anonymous, and the person who wrote the honest comment will work out that you could identify them, which costs you more information than the breakdown was worth.

How to Read the Score

Read the lowest dimension first, the gap between your score and the team average second, and the total last. A strong total with one broken dimension is the most common shape at a small company, and averaging it away is how the broken dimension survives another six months.

Those bands are my working guide rather than a research benchmark, and they are calibrated to the 120-point instrument above. What travels beyond this instrument is the rule underneath them: any single dimension below 12 out of 20 outranks a good total, because five strong dimensions are exactly what conceals a broken sixth.

The founder gap is the second reading and often the more useful one. If you score direction at 19 and the team averages 11, the problem is not that you lack clarity. The problem is that clarity has never left your head in a form anybody else can repeat, which is a distribution failure with a cheap fix: write the three priorities down, put them somewhere permanent, and repeat them at every all-hands meeting until people are tired of hearing them.

The third reading is movement. One round is a snapshot with no context, and the second round six months later is where the instrument starts earning its place. A dimension that moves four points after a specific change tells you the change worked, which is a rare thing to be able to say about anything in this category.

Hard Signals That Corroborate the Score

Five operating numbers you already have will confirm or contradict a health assessment, and they are harder to flatter than a survey. Where the assessment says coordination is weak and the signals agree, you have a finding rather than an opinion.

Only the first of those five has a public baseline. The federal Job Openings and Labor Turnover Survey reported a quits rate of 1.9 percent of employment and a total separations rate of 3.2 percent in July 2026 (Bureau of Labor Statistics). Those are monthly rates, so compare them against your own monthly figure rather than an annual one, and use a consistent turnover calculation so the comparison holds across quarters.

The thresholds in the other four rows are mine rather than published findings, and the trend matters more than the absolute number in every case. Escalation rate is the one I would watch most closely at a growing company, because it rises quietly and it is the earliest visible symptom of decision rights that have not kept up with headcount.

One more signal costs nothing to collect: what happens in the ten minutes after a meeting ends. In a healthy organization, people leave and start work. In an unhealthy one, they schedule a smaller meeting to work out what the first one decided, and that second meeting is where the real cost of poor coordination sits.

What Enterprise Diagnostics Get Right, and What to Leave Behind

Keep the dimension structure and the discipline of scoring the same items repeatedly. Leave the hundred-item instrument, the benchmark database, and the consultant-led readout, all of which exist to solve problems a small business does not have.

The benchmark row is the one worth arguing about. A percentile against other companies is genuinely useful when you have thousands of employees and no other way to know whether a score of 3.8 is good. At twenty people it is close to meaningless, because the variance between two small companies swamps the signal, and your own previous round is a far better comparison than a database of firms nothing like yours.

The instrument length is the other trap. A hundred items produces a beautiful map and a response rate that collapses by round three. A set of 24 statements that gets answered honestly twice a year beats a hundred that get answered once and then abandoned.

Where Organizational Health Assessments Go Wrong

Four failure modes account for most of the wasted effort in this category, and three of them happen after the data is already collected.

The second mistake deserves the extra attention, because it is the one that survives good intentions. Averaging is what an assessment is for at a hundred people and what it does to you at twenty. When five dimensions score 18 and one scores 8, the total of 98 reads as a healthy company, and the company is not healthy. It has one broken system that the other five are quietly compensating for, and the compensation is being paid in hours nobody is counting.

There is a fifth failure worth naming separately, because it is specific to founders. Reading a low score as a personal verdict and reacting defensively ends the exercise permanently, since the team learns that honest answers cost something. Health is a property of the system you built, and rebuilding part of a system is ordinary work rather than an admission about your competence.

What to Fix First

Fix direction before anything else, no matter what the ranking says, unless direction is already scoring above 16. Every other dimension inherits its problems from an unclear direction: decisions are slow because the criteria are missing, handoffs fail because priorities disagree, and renewal stalls because nobody knows which changes would count as improvements.

The direction fix is smaller than it sounds. Write down three priorities for the quarter and one thing you have decided not to do. Put them where people work rather than in a document nobody reopens. Repeat them in one-on-ones until you are bored of your own sentences, which is roughly the point at which the team has heard them enough to repeat them back.

Decision rights come second, and the fix is a list: the five decisions that come up most often, the person who owns each, and the ones that genuinely need you. Most founders discover that two of the five never needed them, and reclaiming those two removes a queue that was slowing down everything behind it.

Coordination is third and the cheapest of the three. End every meeting with a named owner and a date, and make the handoff between two roles explicit about what arrives and in what state. Small companies rarely need process here. They need the two people to have said the same thing out loud once.

Capability is where the mechanical layer earns its place. An explicit team structure, an org chart that says who reports to whom, onboarding a new hire can complete without chasing anyone, and training modules assigned rather than emailed are the things that stop capability from depending on what one person happens to remember. That layer is what FirstHR handles, and it is a deliberately modest claim: the tooling removes the friction, and the six dimensions are still yours to run.

Renewal is last to fix and the one that keeps the rest from decaying. A running assessment is itself a renewal mechanism, provided something visibly changes after each round. That is the whole loop: ask the same 24 questions twice a year, fix one thing, say what you fixed, and let the next round tell you whether it worked.

Frequently Asked Questions