Team Health Check

Purpose: Surface systemic team issues through structured reflection and commit to concrete improvements each quarter

How to run this meeting

Run an anonymous pre-survey, every time. The team health check only works if people say what they actually think — and people don't say what they actually think when their manager is watching them fill out a form. Send a short survey (5–8 questions, 5-point scale plus one open comment field) at least 48 hours before the meeting. Aggregate the results before the session so you're discussing data, not asking people to self-report in real time. Tools like Typeform, Google Forms, or your team's HR platform all work fine for this.

Discuss trends, not snapshots. A single data point tells you almost nothing — a 3.2 on "workload feels sustainable" means something very different if it was a 4.1 last quarter versus a 3.0. The conversation should always be "what changed, and why?" not "is this good or bad?" If this is your team's first health check, set the expectation explicitly that you're establishing a baseline, and the goal of the next check is to see movement.

Focus on systemic issues, not individual complaints. The health check is not a grievance session and it's not an opportunity to surface interpersonal conflicts in front of the team. If someone's open-ended comment clearly refers to a specific person or incident, handle it offline. What you're looking for are patterns: three people independently mentioning unclear priorities, four people flagging that they don't understand how their work connects to company goals, a cluster of comments around on-call burden. Those patterns reveal process and structural problems that you can actually fix. Close the meeting with a commitment to exactly 1–2 changes this quarter — no more. Teams that generate long lists of improvements and act on none of them become more cynical, not less.

Before the meeting

• Send anonymous pre-survey at least 48 hours in advance (5-point scale, open comments)
• Aggregate and anonymize all survey responses before the meeting — never share raw individual responses
• Compute trend lines against previous quarters if data is available
• Identify 2–3 themes from open-ended comments to surface in discussion
• Prepare to share the results honestly, including areas where scores are low — don't soften data
• Invite the full team; this meeting has no observers

Meeting Details

• Date:
• Facilitator:
• Attendees:
• Duration: 60–75 minutes (quarterly)

---

Team Sentiment

Share aggregated survey results across key dimensions. Show trends against prior quarters wherever possible. Let the data lead — don't editorialize before the team has seen it.

Survey: Q1 2026 | Respondents: 9 of 9 (100%)

Overall trend: Consistent decline across clarity, workload, and trust dimensions over three quarters. Recognition is the one bright spot — the peer shoutout process introduced in Q4 is clearly having an effect. The downward trend on decision-making trust is the most significant signal and warrants direct discussion.

---

Workload

Discuss sustainability of current workload honestly. Surface whether pressure is episodic (launch crunch) or structural (permanently understaffed). These require different responses.

Survey average (workload sustainability): 2.7 / 5.0 — lowest score this cycle

Themes from open comments:

• Multiple people mentioned that on-call rotations are coming around too frequently, with not enough time to do "real work" between incidents
• Several comments about context-switching between too many concurrent projects — the team is supporting four different initiatives simultaneously
• At least two comments about weekend pressure, even when it wasn't formally requested

Discussion questions for the room:

• Is the on-call burden the primary driver, or is it the project load?
• Are any of the four concurrent initiatives lower priority than they appear?

Team input captured in meeting: The on-call issue is the loudest signal. Current rotation has 6 engineers, so each person is on-call approximately once every three weeks. With current incident volume, that's unsustainable. The team also flagged that two of the four concurrent projects could be sequenced rather than parallelized without meaningfully impacting timelines.

---

Risks

Identify signals that indicate team health risks if unaddressed — attrition risk, skill gaps, interpersonal friction patterns, or process breakdowns. Be specific about the mechanism of harm.

Risk 1 — Attrition signal (HIGH): The "recommend to a friend" score, while still positive, has declined three consecutive quarters. Combined with low workload sustainability scores and anecdotal conversations @em has had 1:1, there are 2–3 team members who may be passively exploring other options. Sustainable workload is the retention lever most within our control.

Risk 2 — Clarity gap undermining autonomy (MEDIUM): The "understand how my work connects to goals" score has dropped 0.7 points in two quarters. This is the score most correlated with engagement in the research literature. The team is working hard but the connection to company direction feels unclear — this became acute after the Q4 strategy pivot. Engineers are making local decisions without enough context to make good ones.

Risk 3 — Trust in decision-making (MEDIUM): Score dropped from 4.1 to 3.5. Open comments don't name a specific decision, but the timing correlates with the architectural direction change made in January without team input. Worth acknowledging directly.

---

Improvements

Commit to 1–2 specific, actionable changes this quarter. Vague commitments erode trust. Each improvement should have an owner and a way to measure progress at the next health check.

The team discussed six potential improvements and voted. The two selected for this quarter:

Improvement 1 — Restructure on-call rotation

• What: Expand the on-call pool to 8 engineers by training 2 additional team members, reducing rotation frequency from every-3-weeks to every-5-weeks.
• Owner: @em + @sre-lead
• How we'll know it worked: Workload sustainability score improves by 0.5+ points next quarter; on-call rotation frequency documented and visible to the team.
• Target: In place by April 30, 2026

Improvement 2 — Monthly company context session

• What: @em will run a 30-minute monthly session where the team can ask questions directly about company direction and how team priorities connect to it. No prepared slides — just Q&A.
• Owner: @em
• How we'll know it worked: "Connection to company goals" score improves next quarter; attendance tracked.
• Target: First session by March 31, 2026

Improvements not selected this quarter (deferred for consideration in Q2): async retrospective format, dedicated focus time blocks, project sequencing proposal)

---

Action Items

Follow-up

Share an anonymized summary of the results and the two committed improvements with the entire team within 48 hours — this closes the loop and demonstrates that the input was heard. Do not share anything that could identify individual respondents. At the start of the next health check, open by reviewing what was committed last quarter and whether it was delivered. Teams lose faith in this process fast if commitments are forgotten; they build faith when they see evidence that the meeting changes something.