How We Built Teambridg's New Wellbeing Dashboard

Why We Built This

In our 2021 roadmap, we mentioned employee wellbeing metrics as a major focus. Today, I want to pull back the curtain on how we went from idea to shipped feature.

The impetus was clear: burnout reached crisis levels in 2020. Employees working from home blurred the line between work and life, leading to longer hours, fewer breaks, and mounting exhaustion. Managers knew burnout was a problem but had no early warning system — by the time someone said "I'm burned out," it was often too late.

We asked ourselves: could we use the work pattern data Teambridg already collects to identify burnout risk before it becomes a crisis? Turns out, you can — if you know what to look for.

The Research Phase: What Signals Matter?

We partnered with organizational psychologists and reviewed the academic literature on burnout indicators. The Maslach Burnout Inventory — the gold standard in burnout research — identifies three dimensions: emotional exhaustion, depersonalization, and reduced personal accomplishment. We couldn't measure those directly through work patterns, but we could identify behavioral proxies.

After analyzing anonymized, aggregated data from consenting Teambridg customers, we identified five signals that strongly correlate with self-reported burnout:

1. After-hours work frequency — not an occasional late night, but consistent patterns of work outside contracted hours. 2. Declining focus time — a gradual reduction in uninterrupted deep work blocks, suggesting fragmentation and context-switching overload. 3. Weekend activity increases — work creeping into weekends. 4. Communication pattern changes — reduced collaboration frequency, which can indicate withdrawal. 5. Work session length changes — either significantly longer sessions (overwork) or shorter, fragmented sessions (difficulty concentrating).

Designing for Sensitivity

This was the hardest part. Wellbeing data is inherently sensitive. We had heated internal debates about how to present these metrics — and who should see them.

Our design principles were:

No individual wellbeing scores visible to managers. Managers see team-level aggregates and trends. If the team's wellbeing indicators are declining, it's a signal to investigate — not to point fingers. Individuals see their own data. Each employee has access to their personal wellbeing dashboard with specific, actionable insights. "Your after-hours work has increased 40% this month" is a powerful self-awareness tool. Suggestions, not diagnoses. We deliberately avoided language like "burnout detected" or color-coded risk scores. Instead, we use phrases like "pattern shift detected" and "consider reviewing workload distribution."

The Technical Implementation

Under the hood, the wellbeing dashboard uses a rolling 12-week baseline for each user. We compare current-week signals against the individual's own baseline rather than team averages — because what's normal for one person might be unusual for another.

The signal processing is intentionally conservative. We require multiple indicators to shift simultaneously before flagging a pattern change. A single late night doesn't trigger anything. A pattern of late nights combined with declining focus time combined with increased weekend work does. This minimizes false positives while catching genuine shifts.

We also built in seasonal adjustment. We know that certain periods (year-end, product launches, Q4 for finance teams) naturally involve more intensive work. The baseline adapts to account for predictable cycles, so a crunch period doesn't immediately flood the dashboard with alerts.

The feature ships this week for all Teambridg Pro and Enterprise customers. Free-tier users will see a limited version of their personal wellbeing insights. We'll be iterating on this throughout 2021, and we'd love your feedback.