Quiet Quitting: Why Relationships Bind Harder Than Benefits
13 percent of employees in Germany have quietly quit, 77 percent do the bare minimum. What actually halts the withdrawal is not pay or perks, but an internal network that cannot be taken along.
Quiet quitting is not an event, it is a process. Nobody decides on a Tuesday morning to withdraw emotionally. The withdrawal happens gradually, over months, and it is visible long beforehand – just rarely where companies are looking.
This article covers both: what quiet quitting is and how it develops, and why the internal web of relationships is its most effective counterweight. Not for reasons of sympathy, but because it creates switching costs no competitor can outbid.
The situation in Germany
Gallup has run its Engagement Index for Germany since 2001. For the 2025 edition, published in March 2026, 1,700 employees aged 18 and over were surveyed by telephone. The result has been remarkably stable for years, and remarkably poor.
Willingness to move fits the picture: 12 percent of employees are actively looking for a new job, a further 25 percent are open to offers. The gap between the groups is considerable: among those who have quietly quit, 39 percent are actively searching, compared with 3 percent among the strongly attached.
One note on interpretation: the cost estimate is an extrapolation from absence days and productivity assumptions, not a measurement. Critics consider its precision spurious. The order of magnitude is still meaningful – and the 10 percent attachment rate stands as a finding regardless of any cost calculation.
Quiet quitting shows up in behaviour before it shows up in metrics
The most common mistake in dealing with quiet quitting is assuming an annual employee survey will surface it. It does surface it – but late, in aggregate, and without any indication of who is meant.
Observable behaviour is the earlier indicator. Timothy Gardner, Chad Van Iddekinge and Peter Hom identified and validated thirteen so-called pre-quitting behaviours: declining productivity, less team-oriented conduct, reduced initiative, reluctance around long-term commitments, hardly any new ideas. Managers rated their employees' behaviour; actual turnover was recorded thirteen months later. The relationship was clear.
Peter Gloor and colleagues found a similar pattern in communication data. Among 866 managers at a global services firm, network position, response times and communication patterns changed measurably around five months before departure.
Both findings say the same thing: withdrawal leaves traces long before anyone resigns. It shows up first in relationships – in conversations that stop happening, in connections that thin out. Measuring once a year shows you the outcome, not the trajectory.
One caveat belongs here: the Gardner sample contained only 17 actual departures. The qualitative finding is robust, the exact effect size is not. And part of the behaviour Gloor observed likely falls after the decision was already made, since the company studied had a three-month notice period.
The lock-in effect: why relationships bind differently from benefits
Anyone leaving a platform like LinkedIn does not lose the software. They lose their contacts. That is exactly what makes switching expensive, and exactly why people stay on platforms that stopped convincing them long ago.
The same principle applies to employment, and it has a name. In 2001, Terence Mitchell, Brooks Holtom and Thomas Lee introduced the concept of job embeddedness – how firmly a person is anchored in their organisation. They describe three dimensions.
Links – connections
The formal and informal ties to colleagues, teams and committees. Someone connected to many people hangs by more threads.
Fit – compatibility
How well the role, the team and the culture suit the person. Fit develops over time and cannot be created by contract.
Sacrifice – what leaving would cost
Everything that would be forfeited on departure. Not only money, but trust, standing, shortcuts, and people who know you well enough to judge your work.
The meta-analysis by Alex Rubenstein and colleagues covers 29 studies with more than 31,000 employees and reports a relationship between job embeddedness and actual turnover of ρ = –.26. That puts it on a par with job satisfaction and commitment. Its real value lies elsewhere: embeddedness explains retention in addition to satisfaction, commitment and perceived alternatives. It captures something the classic factors do not.
The most important finding is an inconvenient one
The obvious conclusion would be: the more contacts, the better. Research says otherwise. In a 2024 meta-analysis – 22 studies from Thailand and Indonesia covering just over 8,000 employees – the links dimension, the sheer number of connections, is the weakest predictor and not statistically significant. The effect is carried by fit and sacrifice.
Two caveats belong here: the study examines intention to leave rather than actual departures, and it draws on collectivist work cultures. The direction, however, matches findings from other regions.
Cen April Yue and colleagues confirm this from another angle in 2026: the size of someone's professional advice network shows no significant relationship with intention to leave – but the share of genuine friendships within it does.
In practice this means: an internal network binds not because it is large but because it is irreplaceable. Five people who know how you work, whom you can trust and where the shortcuts are, weigh more than fifty entries in the company directory. Networking initiatives that optimise for contact counts are optimising the wrong number.
What the evidence on relationships and retention actually shows
The strongest single study on this comes from Kevin Mossholder, Randall Settoon and Stephanie Henagan. They followed 176 hospital employees over five years, examining how position in the internal communication and advice network affected voluntary departures.
What stands out in the Mossholder finding is the control for job satisfaction. Network position still predicted retention when satisfied and dissatisfied employees were compared. Attachment through relationships is therefore not simply another word for satisfaction.
The meta-analysis by Feeley and colleagues pools five studies on network centrality and turnover and finds r = .29. With five studies that is robust as a direction, not as a point estimate – a caveat worth stating.
Belonging and team spirit
Alongside the switching-cost argument there is a simpler one: people who have colleagues they enjoy working with come to work more willingly. A McKinsey survey of around 5,500 US employees found that better-connected employees were one and a half times more likely to report belonging and one and a half times more likely to report engagement. Only 22 percent said they felt more connected within the company network at all – among frontline employees it was 9 percent, among senior leaders 45.
The BetterUp study on belonging is frequently cited at this point, with 56 percent higher performance and a 50 percent lower turnover risk. Those figures come from a vendor survey, are not peer-reviewed, and rest on self-reported risk rather than observed departures. Usable as illustration, not as evidence.
Where distributed work genuinely plays a part – and where it does not
It is tempting to name hybrid work as the cause. The evidence is more differentiated, and the difference matters.
What is not true
Hybrid work increases turnover
- Nicholas Bloom and colleagues tested it in a randomised trial
- 1,612 employees, assigned by date of birth
- Result: quit rate fell by 33 percent
- Non-managers −40%, women −54%
- No penalty in performance or promotion
What is true
Distributed work siloes the networks
- Yang et al., Nature Human Behaviour, 61,182 employees
- What was studied is the shift to full-time remote work
- Collaboration time with other groups fell by around 25%
- More ties inside the team, fewer bridges beyond it
- Networks became more static: fewer new connections
One precision matters here: Yang and colleagues studied the pandemic shift to full-time remote work, not hybrid models. That hybrid work produces the same effect in weaker form is plausible but not established by this study. The authors do note that the effects are strongly mediated through colleagues – someone sitting in the office still loses connections to those who are not.
The conclusion is not a return-to-office mandate but an insight into the mechanism. Flexible work is attractive to the individual and therefore binds. What gets damaged is what used to emerge on the side: the connections across team and departmental boundaries. Those are precisely the connections that carry knowledge and create switching costs. If you want them, you have to create them rather than hope for the corridor.
What that means in practice is covered in more depth in the article on networking in hybrid teams.
What companies can actually do
Create connections deliberately instead of hoping
Chance encounters have disappeared in distributed organisations without replacement. Formats like coffee roulette or lunch roulette restore them on purpose – low-threshold and without the feel of a programme.
Match across boundaries, not within them
The loss sits in connections across departments, locations and hierarchy levels. Networking inside one's own team happens anyway. That is exactly why formats like skip-level meetings work.
Start early, not at the exit interview
According to a Microsoft analysis of more than 10,000 new hires, new joiners remain markedly less connected than established colleagues even after six months. An onboarding buddy noticeably shortens that phase.
Aim for durability, not contact counts
Recurring encounters with the same people build trust. One-off meet-and-greets produce business cards. That difference decides whether attachment forms.
Make blind spots visible
In organisations of a few hundred people and up, nobody can tell from intuition which areas are well connected and which are isolated. This can be analysed – with limits, see below.
Can this be measured?
Partly, and the honest part of the answer matters more than the enthusiastic part.
Methodologically, organisational network analysis has existed for decades. It reveals where knowledge converges, where bridges between areas are missing and who is isolated. Rob Cross and Adam Grant found across research in more than 300 organisations that 20 to 35 percent of value-added collaboration comes from just 3 to 5 percent of employees – a concentration that never appears on the org chart.
- Communication metadata is sensitive. Germany's Federal Labour Court has ruled that technical systems fall under works council codetermination as soon as they are objectively capable of monitoring conduct or performance. Intent is irrelevant. Analyses of email, chat or calendar data regularly fall into this category.
- A works agreement does not fix an unlawful basis. In 2024 the European Court of Justice clarified that collective agreements must comply with the general provisions of the GDPR. They can specify an existing legal basis but cannot create a new one.
- Self-reporting remains self-reporting. Even anonymous surveys are subject to impression-management bias. This is not an argument against engagement surveys – the Gallup Q12 meta-analysis covers more than three million employees and is methodologically strong. It is an argument against assuming one number a year is enough to steer by.
What works without these problems: metrics from voluntary networking formats, analysed at an aggregated level. How many connections form across organisational boundaries? Which teams are plugged in and which are not? How does that develop over time? This answers the relevant questions without evaluating individuals and without touching the content of communication.
Networking with Workdate
Workdate brings employees together systematically – across departments, locations, hierarchies and generations. The matching takes skills, interests, location and seniority into account, so encounters happen where they hold.
- Continuous matching instead of one-off meet-and-greets, so connections recur and trust can form
- Deliberately cross-departmental, because these are the connections missing in distributed organisations
- Analysis of connectedness at organisational and business-unit level, including blind spots
- Voluntary participation, GDPR-compliant, no analysis of communication content
- No performance evaluation of individuals, but aggregated metrics for HR and leadership
More on the Coffee & Lunch Roulette use case →
More on the Onboarding Networking use case →
Frequently asked questions
What is quiet quitting?
Quiet quitting describes emotional withdrawal from an employer while the employment relationship formally continues. Those affected still complete their tasks but bring no initiative, no engagement and no attachment. Gallup counts 13 percent of employees in Germany in this group, with a further 77 percent showing only weak emotional attachment.
How do you recognise quiet quitting?
Through observable changes in behaviour, not through survey scores. Research has identified thirteen so-called pre-quitting behaviours: declining productivity, less team-oriented conduct, reduced initiative, reluctance around long-term commitments, hardly any new ideas. These signals correlate with resignations that only follow thirteen months later.
Why do relationships bind more strongly than benefits?
Because they cannot be taken along. Job embeddedness research describes attachment across three dimensions: links, fit, and what someone would have to give up when leaving. Salary and benefits can be outbid by a new employer. An internal network grown over years cannot be outbid; it has to be built from scratch.
Does the number of contacts matter?
No, and this is a common misconception. In the research, the sheer number of connections is the weakest factor. What binds people is the quality and irreplaceability of their relationships: trust, fit, and the value of what would be lost. A few solid connections carry more weight than many superficial ones.
Can internal networking be measured in a privacy-compliant way?
Yes, but not without limits. Under German labour court rulings, analysing communication metadata from email, chat or calendar systems regularly requires works council codetermination, regardless of the employer's intent. What is unproblematic are metrics from voluntary formats at an aggregated level that allow no conclusions about individuals and are not used for performance evaluation.
Related topics
Attachment grows between people
See how Workdate connects your employees on purpose – and how the connectedness of your organisation can be made visible without evaluating individuals.