Attrition Risk

Private beta

See attrition risk before someone leaves

A resignation letter is usually the last sign, not the first. Gitrevio looks for changes in engineering activity that may signal disengagement or a higher risk of leaving. This gives managers time to understand what is happening and have a conversation before it becomes a resignation.

Gitrevio Dashboard Risk Reports AI Chat Notifications AA Auwalu Adamu auwalu.adamu@flexiana.com Attrition Risk Assessment — March 2026 47 engineers monitored · 2 high risk · 5 medium risk High risk engineers 0 engineers requires immediate attention Medium risk 5 engineers monitor closely Avg risk score (team) 0.61 up from 0.44 last month Est. departure window 4–8 wks Sarah Chen (top signal) Risk Scores — March 2026 HIGH RISK Sarah Chen (Backend) 0.78 review frequency −45% (8 wks) · cross-team interactions −60% scope narrowing · 1:1 skip rate increasing Est. departure: 4–8 weeks MEDIUM RISK James Park (Frontend) 0.52 velocity stable · code review depth declining meeting attendance sporadic Est. departure: 2–4 months Shapley Factor Decomposition — Sarah Chen Risk score: 0.78 HIGH engagement decline 38% reduced scope 27% social withdrawal 22% schedule pattern change 13% Suggested intervention — talk before it becomes a resignation: → 1:1 conversation re: growth trajectory → New project assignment (addresses top 2 factors) Impact if leaves: −38% review capacity · 3 orphaned services

How it works

PATTERN DETECTION Sarah Chen · Backend 8-week activity vs. her normal pattern Normal pattern Actual activity W1 W2 W3 W4 W5 W6 W7 W8 Meaningful change detected Risk score 0.78 Est. departure 4–8 wks Flagged weeks before the earliest departure estimate — time for the conversation.

Gitrevio compares each person's current activity with their own normal pattern and looks for meaningful changes over time.

High Risk

Sarah Chen, Backend

Risk score: 0.78

What changed

  • 45% fewer reviews over 8 weeks
  • Working on a narrower set of tasks
  • 60% fewer cross-team interactions
  • More missed 1:1s

Main factors

  • Engagement decline: 38%
  • Reduced scope: 27%
  • Social withdrawal: 22%
  • Schedule change: 13%

Estimated departure window: 4 to 8 weeks

Medium Risk

James Park, Frontend

Risk score: 0.52

What changed

  • Review depth is declining
  • Meeting attendance is less consistent
  • Overall delivery remains stable

Estimated departure window: 2 to 4 months

The score is based on a pattern, not one signal. A drop in commits alone does not mean someone is likely to leave.

Gitrevio also uses each person's own baseline. A change that matters for one engineer may be normal for another.

Risk scores update as new data comes in, so managers can see when a pattern is changing.

Signal Categories

What Gitrevio looks at

Gitrevio looks at changes in engineering activity and compares them with each person's normal pattern.

Commit activity

Changes in commit frequency over time.

Code reviews

Fewer reviews, comments, or review requests.

Work patterns

Changes in working hours or weekend activity.

Time off

Changes in vacation and time-off patterns.

Team activity

Changes compared with the person's team.

Scope of work

Working across fewer projects, repos, or types of work.

Team interaction

Less participation across teams and fewer interactions.

Learning and mentorship

Less involvement in new areas of the codebase, helping others, or onboarding.

Know what may be driving the risk

A risk score tells you that something changed. Gitrevio also shows which signals are contributing to the score.

For example, a high score may be linked mainly to lower engagement and a narrower scope of work.

That gives managers a better starting point for the conversation.

The goal is not to guess why someone wants to leave. It is to spot meaningful changes early and understand them before making a decision.

Sarah Chen · Backend

Risk score: 0.78

Engagement decline 38%
Reduced scope 27%
Social withdrawal 22%
Schedule change 13%

The two largest factors point to a specific conversation — not a generic check-in.

Privacy & Ethics

Built for better conversations, not surveillance

Gitrevio is designed to help managers understand team health and act early.

Only managers and above see individual risk scores.

Engineers can see their own profile and understand what data is being used.

No keystroke tracking. No screenshots. No hour tracking.

Gitrevio uses engineering activity such as commits, pull requests, and code reviews. It works with data teams already create as part of their work.

Early Warning

What you can do with an early warning

01

Start a conversation

Talk to someone before concerns become a decision to leave.

02

Change the work

Move someone to projects that better match their interests and strengths.

03

Offer support

Add mentorship or help where someone may be struggling.

04

Create a growth path

Discuss what's next and what they want to work toward.

05

Review compensation

Check whether pay may be part of the problem.

06

Reshape the team

Adjust responsibilities when the current setup is not working.

Built with responsible analysis

Gitrevio checks the quality of its attrition analysis before showing a risk score.

The analysis considers data quality, model accuracy, calibration, and differences across groups.

Explore people and knowledge tools

Detect attrition risk weeks before it becomes a resignation.

Get started free