What the Second Health Check Tells You

The quarterly health check results are in. The head of people opens the deck. Company-wide morale is at 6.2. Someone at the table says: same as last quarter. There's a beat. Then someone asks the obvious question — is that good news?

Nobody answers. A few people look at the number. Someone says "stable, I suppose." The conversation moves to whether specific teams need follow-up. The meeting ends without anyone knowing whether the organization is in a better position than it was ninety days ago or a worse one. The score is the same. What the score means — whether it represents a floor, a plateau, or a ceiling the organization is pressing against — is not in the deck.

They only have the state. They don't have the direction.

---

What a Score Measures

A score is a snapshot. At the moment the questions are answered and the averages are calculated, the score captures where things stand. That is all it captures.

The problem is not that 6.2 is an imprecise number. The problem is that 6.2 is a complete description of a state and an incomplete description of a situation. A situation has movement. A state does not. And movement is what produces decisions.

Consider the difference in what a decision-maker actually needs. A team at 6.2 and falling — 7.4 three quarters ago, 6.8 two quarters ago, 6.2 now — is in a different situation from a team at 6.2 and climbing — 5.1 twelve months ago, 5.7 six months ago, 6.2 now. The number is identical. The situation is opposite. One is a crisis signal with momentum; the other is a recovery with momentum. The response appropriate to each is entirely different. Not slightly different — different in kind.

A score without direction cannot distinguish between those two situations. It doesn't matter how accurately the score was measured or how carefully the questions were designed. A single data point cannot be a trend. A trend requires at least two points, which means it requires at least two instances, which means it requires running the same exercise more than once.

What most organizations do is run the health check and act on the score. What they should do is run the health check and wait for the second one.

---

Why One Health Check Can't Tell You

The language used around health checks implies that the exercise produces intelligence. "We ran our health check." "We have the results." But what was produced is a data point — a single observation. A data point is not intelligence. It is the raw material from which intelligence might eventually be constructed.

This distinction matters because it determines what you can do with the output.

A data point tells you where something is. Intelligence tells you where something is going. You can acknowledge a data point. You can respond to it emotionally, express concern, commit to follow-up. What you cannot do is make a decision about direction — because direction requires comparison, and a single instance has nothing to compare against.

The Head of People who presents quarterly health check results is presenting a state. The leadership team receives it as one. They acknowledge it, note the teams below threshold, commit to conversations. And then they wait three months for the next score, at which point they will compare the two scores manually, probably in a spreadsheet, probably imperfectly, and probably after the fact when it's already obvious that something has moved.

The structural issue is that a single instance is not designed to accumulate. Each health check sits alone. The questions may be the same, the cadence may be quarterly, but without a system that links instances — that treats each run as one observation in a longitudinal dataset rather than a standalone exercise — the comparison never becomes automatic. The direction is always an afterthought, never the first thing visible.

This is where the timing problem compounds the structural one. The manual comparison — pulling last quarter's deck, lining up the numbers — happens in the meeting room, at the moment someone thinks to ask. By then, the question is reactive. The situation has already developed. The signal that a score was moving was present in the data weeks earlier, but no one saw it because the data from each instance was sitting in its own file, unconnected to what came before. What should have been a leading indicator arrives as a lagging one.

A dataset produces something a data point cannot: the signal that a situation is changing before the change has become obvious. The difference in value between a data point and a dataset is not incremental. The direction is only visible in the dataset — and direction is what the decision requires.

---

What Running It Twice Actually Gives You

When the same interview is run on a regular cadence and the instances are linked structurally, the output of the second run is different in kind from the output of the first.

The first run produces a state. The second run produces a state and a delta. The delta is the first piece of real information. A morale score of 6.2 tells you something is at a level. A delta of +0.4 over one quarter tells you the situation is recovering. A delta of −1.2 over three quarters tells you there is a deteriorating situation that has been declining for the duration and that the current number understates the severity of the trajectory.

That delta — not the score — is what determines the category of response. A recovering situation requires sustaining whatever changed. A deteriorating situation requires finding what changed. A flat line over multiple periods requires examining whether the score reflects stability or stagnation. None of those analyses are possible from a single observation. All of them are immediate from a dataset of four.

The Head of People who runs the same session quarterly and compares instances structurally does not need to wait for the quarterly meeting to know whether a team is in trouble. The trajectory is visible as data comes in. A team that was at 7.4 and is now at 6.2 is sending a different signal before anyone in the room has named it. The question "is that good news?" has an answer before the meeting starts.

Series trend analysis produces not just themes from the content of the sessions but directions for each: Improving, Declining, Stable, Emerging, Resolved. Those five categories are answers to the question that otherwise never gets answered in the meeting room. They convert a state into a situation. They make the delta the headline, not the footnote.

The head of people who walks into the leadership review with a series trend report is not presenting a score for the room to interpret. She is presenting a direction — with evidence — that names its own required response.

---

The One Thing the Score Can't Show You

The quarterly health check will keep coming back. The scores will be acknowledged, the commitments will be made, the follow-up conversations will happen. And in ninety days, the same scores may return, and the same question will be asked — is that good news? — and the answer will be the same non-answer.

Not because the organization doesn't care. Because the output of a single health check instance is structurally incapable of answering the question the room is actually asking. The question is not "where are we?" The question is "are we moving?"

A score without a direction is not information. It is a measurement. The direction is what the decision requires — and the direction only appears when the same questions are asked, on the same cadence, and compared against what came before.

The second health check doesn't just give you a second score. It gives you the first piece of intelligence.

---

Mayetik helps heads of people run health checks as structured recurring sessions, tracks each instance in a series, and surfaces trend direction — Improving, Declining, Stable — so the quarterly review starts with the delta, not the question of whether anyone knows what the number means. Start your free trial.