Loose Coupling: Why HR Policy and Actual Behavior Live in Different Buildings

60-Second Summary

Weick (1976): organizations are loosely coupled systems where subunits are responsive to each other but retain their own identity and separateness.
The policy layer and the practice layer are only loosely coupled. Your leave policy and how leave is actually granted are two systems, not one.
Benefits: local adaptation, resilience, room for experimentation. Costs: inconsistent experience, policy-practice gaps, bias hiding under formal fairness.
The fix isn't tight coupling — that's brittle. It's making the loose couplings visible and monitoring the gap.
The best HR leaders know which policies are tightly coupled to practice, which are loose, and which are theatre.

The handbook offers 16 weeks of paid parental leave. Uptake data shows the average male employee takes 3. The policy is generous. The practice is not. That is loose coupling — and it's usually where the real story lives.

What Weick actually meant

“Loose coupling connotes situations in which entities affect each other suddenly rather than continuously, occasionally rather than constantly, negligibly rather than significantly, indirectly rather than directly, and eventually rather than immediately.”
— Karl Weick, Administrative Science Quarterly (1976)

Weick argued that many organizations aren't tightly integrated machines — they're collections of subsystems held together by weak, occasional connections. Policy doesn't fully determine practice; org charts don't fully determine reporting reality; formal decisions don't fully determine what actually happens on Monday. This is not dysfunction — it's how organizations absorb change without shattering.

Loose coupling across HR

Every row is a place where policy and practice are loosely coupled — sometimes usefully, sometimes harmfully.

Domain	Policy layer	Practice layer	Typical gap
Parental leave	16 weeks paid, all parents	Fathers take 3 weeks average	Culture, manager signalling, career fear
Performance reviews	Calibrated rubric, evidence-based	Manager wrote it in 20 min the night before	Time, training, workload
Promotion criteria	Ladder with anchors	Who advocated hardest in committee	Sponsor advantage, visibility bias
Hiring rubric	Structured scoring, evidence quotes	Interviewer decided in minute 4	System 1 dominance
Code of conduct	Zero tolerance for harassment	Repeated behaviours quietly tolerated	Escalation cost, reputation risk
Wellbeing / mental health	EAP available, therapy covered	Uptake very low, stigma high	Top-down signalling, privacy fears

Managing the gap honestly

Denial vs deliberate management

Denial (typical)

'The policy says X, so X is what happens'
Uptake and outcome data not monitored
Handbook edits treated as behaviour change
Manager variation ignored in reporting
Complaints treated as isolated cases

Deliberate management

Policy and practice tracked separately
Uptake, variance, and outcome data reported
Behaviour change treated as its own project
Manager-level variance analysed and coached
Patterns across complaints trigger design review

Five moves to make the coupling honest

1
Instrument the practice layer
Measure not just policy existence but usage: leave uptake by gender/manager, promotion rate by demographic, PIP outcomes by manager.
2
Publish gaps internally
'Our policy is 16 weeks; our male uptake is 3. Here's why we think that gap exists and what we're doing about it.' Naming dissolves half its power.
3
Coach the loose points, not the policy
Rewriting the handbook rarely changes behaviour. Coaching the manager, calibration room, and specific decision does.
4
Accept useful looseness
Local adaptation, humane exceptions, and pragmatic workarounds are often features. Tight coupling everywhere is bureaucratic misery.
5
Tighten where safety matters
Harassment, discrimination, pay, and safety warrant tight coupling. Elsewhere, tolerate looseness deliberately.

FAQ

Frequently asked questions

Isn't the goal to eliminate the gap?

No. Tightly coupled organizations are brittle. The goal is deliberate looseness — knowing which gaps are features and which are failures.

How is this different from 'culture eats strategy'?

It's the mechanism. That aphorism is the outcome; loose coupling is the structural explanation.

Where does this go wrong most badly?

When leadership uses the policy layer as PR while the practice layer contradicts it. Employees notice; the trust deficit compounds.

Takeaways

Policy and practice are two layers, loosely coupled — always. The question is whether you manage the gap.
Instrument the practice layer. Metrics on usage, variance, and outcomes are the coupling.
Publish gaps internally. Naming them creates accountability.
Not all looseness is bad. Choose deliberately where to tighten.