Gallup research finds 70% of team engagement is attributable to the manager. But most managers receive little guidance on how to actually read their team's productivity signals -- what to watch, what deviation means, and when to act. This guide provides that framework using data from tools teams already use.
- Focus on output velocity and behavioral patterns, not activity metrics -- time online is not a proxy for productive work
- The signal that matters is deviation from an individual's own baseline, not comparison to team averages
- Data prompts the conversation; the conversation surfaces the cause -- never use behavioral data to draw conclusions without a human dialogue
Gallup's 2025 research delivers a number that should matter deeply to anyone in organizational management: 70% of team engagement is attributable to the manager. Not the company culture. Not the compensation package. Not the office perks. The manager.
This makes the current state of manager support troubling. Gallup found that manager engagement fell globally in 2024, and only 44% of supervisors received formal management training. The people who most influence team health are the least equipped to do so systematically.
Part of the gap is tool-related. Managers have access to enormous amounts of data about how their teams work -- in Slack, in Jira, in GitHub, in calendar systems -- but no framework for reading it. This guide provides that framework.
The Core Principle: Baselines and Deviation
Every signal framework depends on a baseline. You cannot know that something has changed without knowing what normal looks like. Before trying to read your team's productivity signals, you need to establish individual baselines during a period of normal operations.
What to baseline for each person:
- Average response time in team messaging during core hours (measure over 4-6 weeks)
- Output velocity: tasks/tickets closed per week, average across a normal sprint or project cycle
- Calendar availability pattern: what blocks are typically free versus scheduled
- Meeting participation norm: camera usage, response latency in calls, verbal contribution frequency
- Proactive vs. reactive communication ratio: do they initiate discussions, or only respond?
With baselines established, deviations become meaningful. Without them, you are just observing snapshots without context.
Reading the Signal Sources
From Your Project Tracker (Jira, Linear, Asana)
Project trackers are the most direct source of output signal. What they tell you:
The most underused metric in project trackers is velocity consistency. Managers watch for average output, but the consistency of that output is often more informative. An employee delivering 8 points one week and 3 the next, repeating, is showing a different signal than someone delivering 5-6 consistently.
From Your Communication Tool (Slack, Teams)
Communication tools surface behavioral patterns that project trackers cannot. Key signals:
- Response time distribution: Not just average, but variance. High variance (sometimes very fast, sometimes very slow) indicates competing demands. Sustained high average indicates disengagement or overload.
- Proactive message ratio: Employees who initiate discussions are more engaged than those who only respond. A shift from proactive to purely reactive communication is meaningful.
- Channel breadth: Engaged employees participate across multiple relevant channels. An employee retreating to only their direct work channels is narrowing their team investment.
From Calendar Data
Calendar availability reveals scheduling constraints that are invisible elsewhere. A sudden increase in recurring blocks that cannot be attributed to any visible team project, or a pattern of consistent unavailability in specific time windows, is worth a check-in.
The most valuable insights often come from combining quantitative tracking data with qualitative feedback from team members about their experience. (Slack research on productivity tracking)
The Intervention Framework
Reading signals only matters if it prompts appropriate action. A tiered response:
Level 1: Mild, early signal (1-2 signals, less than 2 weeks)
No action required yet. Continue observing. Note the observation in a private log with specific dates and data points. This is your baseline for determining whether the signal is sustained.
Level 2: Cluster or sustained signal (2+ signals, 3+ weeks)
Schedule a genuine check-in. Not a performance conversation -- a human one. "I wanted to check in on how things are going. I've noticed your response times have been a bit slower lately and I wanted to make sure nothing is on your plate that I'm not aware of." The conversation may resolve everything with a simple explanation, or it may surface something that needs attention.
Level 3: Strong cluster, sustained, with performance impact
Involve HR. Document the behavioral pattern with specific dates and data. Prepare for a more structured conversation with a clear performance expectation and timeline. Continue the human dialogue -- understanding root cause matters for resolution.
What Not to Do
A few common mistakes that undermine the value of signal-based management:
- Do not use signals to draw conclusions without conversation. Every signal has multiple possible explanations. Use the signal to prompt dialogue, not to reach a verdict.
- Do not compare individuals to team averages. Different people work differently. A naturally slower responder who maintains their response time is healthy. A fast responder slowing down is showing a signal. Individual baseline deviation is what matters.
- Do not over-index on any single metric. Response time alone is not diagnostic. Output velocity alone is not diagnostic. It is the cluster and the pattern over time that creates a meaningful signal.
- Do not make this feel like surveillance. If your team knows you are reading behavioral signals to support them -- not to catch them -- the data serves a healthy management function. Transparency about what you observe and why matters.
Related reading: Leading vs lagging indicators in remote team health and warning signs your best employee is about to quit.
See what your tools already know
Signal does the baseline-building and deviation-detection automatically, delivering weekly AI-synthesized digests that tell you which signals appeared and what they might mean.
Explore Signal →