Why Health Check Data Sits in a Deck and Changes Nothing

The health check results come back in February. The HR leader presents the deck in the leadership review: company-wide engagement at 6.8, up half a point from last quarter. One team — engineering — is at 4.8, down from 5.1. The room is sympathetic. Someone says they'll work with the engineering manager to understand what's going on. Another suggests a team offsite. The meeting moves to the roadmap conversation. The deck is filed.

In August, the health check results come back again. Engineering is at 4.7.

This is not a story about leadership that doesn't care. The follow-up happened — there was a 1-on-1, a team lunch, a brief offsite. The conversation in February was genuine. What it didn't produce was a change, because neither the conversation nor the follow-up was grounded in a diagnosis.

A score is not a diagnosis. And a conversation built on a score produces responses appropriate to a score — acknowledgment, concern, a commitment to look into it — not responses appropriate to whatever is actually happening.

What the Score Doesn't Contain

A health check score tells you there's a problem. It tells you nothing about what kind.

"Engineering is at 4.8" could mean a specific manager relationship is generating consistent friction. It could mean the team is absorbing the aftermath of the Q2 incident and the technical debt that followed, and the score reflects exhaustion, not dissatisfaction with the organization. It could mean the team has been cut out of technical decision-making in the months after the incident, and the score reflects disengagement from a process they used to be part of. These are three entirely different problems. The responses appropriate to each are entirely different. None of them are visible in the score.

The aggregate score is the point at which the diagnosis disappears. The health check collected the signal — in the experiences of the people who answered it — and then averaged those experiences into a number that no longer contains them. Leadership is left with a number that confirms something is wrong and offers no direction for what to do about it. The follow-up conversation fills the gap with hypotheses. Someone who knows the engineering team offers a theory. Someone else offers a different one. The conversation is about figuring out what the number means, and it ends without a clear answer because the number, structurally, cannot provide one.

The Question the Score Can't Answer

Underneath every health check review is a question that the score cannot answer: is this problem isolated or systemic?

Isolated means one team's situation, with a specific cause traceable to that team's experience. Engineering's score reflects a specific event or dynamic that engineering has lived through and other teams haven't. The response to an isolated problem is targeted — address that team, address that cause.

Systemic means the same concern appearing across multiple teams independently, pointing at something upstream of any single team's experience. If engineering describes a feeling of unclear strategic direction, and product describes the same feeling in different language, and go-to-market describes it in a third formulation, that isn't three separate team problems. It's one organizational pattern expressing itself in three places. The response to a systemic problem is different in kind: it isn't a team-level intervention. It's a leadership behavior or organizational decision that's generating the same signal in multiple places.

The health check as commonly run — aggregate score, leadership review, follow-up conversation — cannot make this distinction. To separate isolated from systemic, you need accounts collected across teams independently, synthesized, and examined for what appears in one team's account versus what appears in three independently. The score cannot tell you that. The deck cannot tell you that.

What the Synthesis Finds

A structured health check run as sessions across all departments — accounts captured independently — produces a synthesis that makes the isolated/systemic distinction visible.

Isolated problems have a specific signature in the synthesis: the language and the cause appear in one team's account and nowhere else. Engineering describes frustration with the technical decision-making process in the months after the incident. No other team describes anything similar. That's engineering's problem — specific, traceable, addressable at the team level.

Systemic problems have a different signature: the same theme appearing in multiple accounts independently, often in different language. Product describes feeling out of the loop on strategic decisions. Go-to-market describes a sense that the company's direction shifts more often than it used to. Infrastructure describes executing on priorities that change before the work is finished. Three formulations of the same underlying experience — and none of those teams were in the same room when they described it.

The synthesis is what makes the pattern visible. Without it, leadership sees a 4.8 in engineering, a 5.4 in product, and a 5.6 in go-to-market and reads three separate numbers. With it, they see that two of those three numbers are expressing the same concern — strategic clarity — that no single team's intervention can address. The isolated problem in engineering is real and separate. The systemic pattern across product, go-to-market, and infrastructure is also real, also separate, and requires a completely different response.

Both would have stayed invisible in the deck.

What Changes for the HR Leader

The HR leader who presents scores offers data for interpretation. The HR leader who presents a synthesis offers a diagnosis that has a required response.

"Engagement is at 6.8, with engineering at 4.8" invites interpretation. The room fills with hypotheses. Someone offers a theory about the engineering manager. Someone else offers a theory about the incident. The conversation is about figuring out what the number means, and it rarely concludes with a clear decision because the number, by design, doesn't name one.

A synthesis produces a different kind of statement: "Engineering's score reflects an isolated issue traceable to how technical decisions were made after the incident — we heard this consistently in their account and nowhere else. Separately, we're seeing the same concern about strategic clarity appearing in product, go-to-market, and infrastructure independently. That's not a team-level problem. It's something at the leadership level generating the same signal in three different places."

The isolated problem has a tractable cause and an implied owner. The systemic pattern has an implied subject: what leadership decision or behavior is producing this?

The synthesis changes the HR leader's position in the conversation — from presenting a number for the room to interpret, to surfacing a pattern that names its own required response. The score gives leadership permission to acknowledge and move on. The synthesis doesn't offer that option.

The Health Check That Produces a Decision

Most leadership teams genuinely intend to act — the follow-up happens, the 1-on-1 gets scheduled. What doesn't happen is a decision that addresses the actual cause, because the actual cause was never surfaced. The score told the leadership team there was a problem. It didn't tell them whether the problem was that team's alone or something upstream of all of them. It didn't tell them whether addressing engineering's specific situation would change the score in August, or whether the same 4.7 would come back because the source was something else entirely.

Isolated problems can be acted on at the team level. Systemic problems can only be addressed at the level where they originate — and a leadership team that doesn't know a problem is systemic will keep addressing it at the wrong level, and keep getting the same deck in August.

The health check that sits in a deck and changes nothing isn't a data problem. The data was collected. It was never structured to produce the one distinction that makes a response possible: whether what's happening in one team is that team's problem, or everyone's problem showing up in one team first.