7 pre-resignation signals your team is ignoring
Written by Víctor Ciudad-Fernández.
Most key-person resignations are not a surprise: they announce themselves 3-6 months in advance through behavioral and network signals that annual surveys do not capture. At Beetrics we have identified seven recurring patterns that appear in cases of key figures (knowledge holders, informal leaders, talent the organization would want to retain) who end up resigning. The useful intervention window sits between month 4 and month 2 before the declared decision.
This isn't a list of generic warning signs. These are the seven indicators we cross with the organization's real network when a client asks us to identify who's likely to leave in the next six months. A single signal says nothing on its own. Several at once, in one critical person, do.
The 7 signals we keep seeing
1. Progressive drop in cross-team collaboration
The person stops participating in discussions from other areas, stops offering help to neighboring teams and restricts their internal network to their immediate circle. It is the earliest signal, visible 4-6 months in advance.
2. Decline in 1:1 frequency and duration with the manager
Meetings get shorter, get postponed or shift to "cancelled, not rescheduled". The person puts their energy outside the team, and you notice it in their mood before you see it in the numbers.
3. Increased LinkedIn activity outside work hours
Not new, but still the most correlated behavioral indicator: more updates, more connections and more likes on competitor content. Typically appears 2-3 months before the declared resignation.
4. Reduced impact in collective decisions
The person stops pushing opinions in committees, accepts more than they propose and turns "neutral" in discussions where they used to take a stance. It is a silent disengagement that precedes the exit.
5. Mentoring that cools down
The people they were mentoring report fewer sessions, less deep feedback and a shift toward more operational, less strategic answers. Knowledge stops transferring.
6. Lower presence in informal channels
Less participation in coffee chats, less attendance at optional activities, fewer comments in non-critical threads. Belonging evaporates before the contract does.
7. Atypical vacation and time-off patterns
Occasional day requests with no clear pattern, vacations matching external recruiting processes, absences on dates where they used to actively participate. Not proof on its own, but combined with other signals it is highly predictive.
Beetrics rule: a single signal is noise, two combined signals in 4 weeks are attention, three or more in a critical profile are immediate intervention.
Why annual surveys do not detect them
Climate surveys measure how people feel at the moment they answer. They capture declared satisfaction, not actual behavior. People planning to leave are, paradoxically, the worst respondents to a survey: they minimize, they avoid confrontation, they mark "neutral" to not draw attention. That is why reported engagement and voluntary turnover have been decorrelated for years in sector studies like Gallup's.
The signals that matter are elsewhere: in how relationships change, not in how feelings are reported. That is the space covered by organizational network analysis, and it is the difference between a classic engagement platform and a real early-detection tool.
How Beetrics helps when the pattern appears
Effective intervention is not a motivational conversation or a counter-offer. What works, based on the data we keep accumulating at Beetrics, is acting on three fronts at once:
- Load and network: redistribute the dependencies that fall on them, take weight off if they're holding up half the team, give them room to work with the people who still energize them.
- Visible trajectory: make the next career step explicit on a 12-18 month horizon with manager commitment, not an abstract promise.
- Institutional recognition: change in visibility (presence in committees, initiative ownership), not a generic thank-you email.
If the signal appears late (1-2 months in advance), retention probability drops to 30-40% and the focus must shift to managing the transition: documenting critical knowledge, identifying the natural successor in the network, preparing the relationship reorganization the exit will break. Better to arrive on time to an orderly exit than late to a useless counter-offer.
This applies differently by sector. In industrial plant, the most useful signals are absenteeism and changes in the relationship with supervisors; in tech, LinkedIn and reduced collaboration in code reviews weigh more; in hotel chains, the relevant signals are requests for inter-property moves and the decline of mentoring to new seasonal hires.
Frequently asked questions
How far in advance can these signals be detected before a resignation?
The earliest network signals (drop in cross-team collaboration, mentoring cooling down) are visible 4-6 months in advance. The more explicit behavioral signals (LinkedIn, atypical vacations) typically appear 2-3 months before. The useful intervention window is between month 4 and month 2 prior to the declared decision.
How many signals does it take to be concerning?
A single signal is noise. Two or three combined signals in a key profile sustained for 4-6 weeks constitute a pattern. Beetrics applies a combined network + behavioral score that weights profile criticality and intensity of change, not the binary presence of each signal.
How is this measured without invading employee privacy?
Beetrics works with aggregated, opt-in data. Network signals come from brief surveys where each employee declares who they collaborate with; behavioral signals are only incorporated if the client connects anonymized sources. It complies with GDPR and AEPD by design: no individual emails or messages are observed.
What if the person has already made the decision?
If signals appear in their late phase (1-2 months), retention probability drops to 30-40%. Even so, intervening at that point allows managing the transition: documenting critical knowledge, identifying the natural successor in the network, preparing the relationship reorganization the exit will break.
Does this replace exit interviews?
No, it is the temporal opposite. Exit interviews explain why the person left, and there is no margin left to act. Pre-resignation signals let you act before the decision is irreversible. Both are complementary.
Does it work the same in industrial plant as in tech teams?
The conceptual signals are the same (disengagement, disconnection, external search), but their form changes. In plant we see more absenteeism patterns and changes in relationships with supervisors; in tech we see more LinkedIn activity, less collaboration in PRs and code reviews, less attendance at all-hands. Beetrics adapts the model per sector.
Want to see which signals are already showing up in your organization?
Request a demo