14.3 Systems Theories in OD: Open Systems, Complexity, Network Theory & Action Research
Key Takeaways
- Open systems theory treats the organization as importing inputs, transforming them, exporting outputs, and adapting through feedback, which is why an intervention inside one subsystem is absorbed or reversed by the subsystems around it.
- Requisite variety holds that a control system must be at least as varied as the environment it regulates, which explains why a single standardized programme fails across genuinely diverse operating contexts.
- Complexity theory distinguishes complicated problems, which yield to expert analysis, from complex ones, which require safe-to-fail probes because cause and effect are only visible in retrospect.
- Network theory reveals the informal organization, where brokers spanning otherwise disconnected groups carry disproportionate influence over whether a practice spreads, regardless of formal seniority.
- Action research is the participatory cycle of diagnosis, action, and evaluation with the client system, and it is the methodological foundation on which organization development rests.
Why Systems Frames Matter to a Talent Development Practitioner
The content outline asks for knowledge of frameworks describing the design, interaction, and operation of social, organizational, and informational systems. The practical value is diagnostic: each frame makes a different class of problem visible, and reaching for the wrong one produces a confident analysis of the wrong thing. This section covers the five named frames and the choice between them.
1. Open Systems Theory
Derived from von Bertalanffy's general systems theory and applied to organizations by Katz and Kahn, open systems theory holds that an organization is not a closed machine but a system that imports energy and information from its environment, transforms it, exports outputs, and adjusts through feedback.
Key properties the exam rewards:
- Input - transformation - output - feedback. Every subsystem sits in this cycle, and a change to one node propagates.
- Interdependence. Subsystems are coupled, so an intervention in one is absorbed, amplified, or reversed by the others. This is the theoretical reason that a training intervention alone rarely changes a performance outcome: the reward system, the workflow, and the supervisory practice all continue to push the original behaviour.
- Equifinality. A system can reach the same end state from different starting conditions and by different routes. There is no single correct organizational design, which is why copying another company's operating model reliably disappoints.
- Homeostasis. Systems act to restore a steady state. Improvement that is not embedded structurally decays back toward the prior equilibrium, which is the systems explanation for why pilots succeed and scaled rollouts regress.
- Requisite variety (Ashby). A regulating system must possess at least as much variety as the system it regulates. A single standardized curriculum imposed across genuinely different operating contexts lacks the variety to control them, and local workarounds emerge to supply the missing variety.
2. Chaos and Complexity Theory
Complexity theory addresses systems whose behaviour emerges from interactions between many agents rather than from a central design. Its most useful contribution to practitioners is the distinction between problem types:
| Domain | Relationship between cause and effect | Right approach |
|---|---|---|
| Simple / clear | Obvious to everyone | Apply best practice |
| Complicated | Discoverable through analysis by experts | Analyse, then apply good practice |
| Complex | Visible only in retrospect | Run safe-to-fail probes, amplify what works, damp what does not |
| Chaotic | No discernible relationship | Act to establish stability first, then move the situation into another domain |
Related concepts: emergence, where system-level patterns such as culture arise from local interactions and cannot be installed by decree; sensitivity to initial conditions, where small differences in how a change begins produce large divergence later, which is why the first cohort of a rollout matters disproportionately; and attractors, the stable patterns a system keeps returning to.
The practitioner's error to avoid is treating a complex problem as complicated — commissioning a detailed analysis and a comprehensive plan for a situation whose causal structure will only be legible afterwards. Culture change, cross-functional collaboration, and engagement are complex; a payroll integration is complicated.
3. Network Theory
Network theory maps the informal organization: who actually goes to whom for advice, information, and help, which frequently bears little relation to the org chart. Organizational network analysis surfaces four roles:
- Central connectors — high in-degree; many people rely on them. They accelerate adoption and become bottlenecks.
- Brokers (boundary spanners) — they connect otherwise disconnected clusters. Their influence is structural rather than hierarchical, and because they carry information between groups that do not otherwise talk, they determine whether a practice crosses functional boundaries at all.
- Peripheral actors — weakly connected; often new, remote, or disengaged, and the first to leave.
- Energizers — interactions with them raise others' motivation, which is a distinct property from connection count.
Two findings matter for talent development. First, weak ties carry novel information — Granovetter's result that new opportunities and ideas travel along infrequent cross-cluster ties rather than within close-knit groups. Second, structure beats seniority for diffusion: identifying and engaging brokers spreads a practice faster than a cascade down the formal hierarchy, because the hierarchy is exactly the structure the practice is failing to cross.
4. Action Research
Action research, originating with Kurt Lewin, is the methodological foundation of organization development. It is participatory and cyclical: the practitioner works with the client system rather than on it, and diagnosis, action, and evaluation repeat rather than running once.
The cycle: enter and contract - collect data - feed data back to the client system - jointly plan action - take action - evaluate - re-enter the cycle.
Two features distinguish it from conventional consulting. First, the data feedback step is itself an intervention: showing a group its own data changes the group, which is why feedback design matters as much as data quality. Second, the client system participates in diagnosis, which is what generates the ownership that makes action stick — a diagnosis delivered to a client produces a report; a diagnosis reached with a client produces commitment.
Choosing a Frame
| Presenting problem | Most useful frame |
|---|---|
| An intervention worked in pilot but decayed at scale | Open systems — homeostasis and subsystem interdependence |
| A standardized programme is being locally worked around | Open systems — requisite variety |
| Comprehensive plans keep being overtaken by events | Complexity — treat as complex, run safe-to-fail probes |
| A practice is not crossing functional boundaries | Network theory — find and engage the brokers |
| The client resists a diagnosis they did not participate in | Action research — rebuild the engagement participatively |
Exam Trap: A distractor that responds to a complex, emergent problem with a more detailed analysis and a more comprehensive implementation plan is applying a complicated-domain method to a complex-domain problem. The systems-literate answer runs small parallel probes and amplifies what works rather than committing to a single large plan.
A global manufacturer pilots a continuous-improvement programme at one plant with excellent results: defect rates fall 31% and the practice is sustained for nine months. The programme is rolled out unchanged to eleven other plants. Within two quarters, ten of the eleven have reverted to prior defect rates, and local teams describe the standard method as 'not how things work here.' Which systems reading best explains this and points to the right corrective action?
A talent development team wants a new peer-review practice to spread from the engineering function into product management and customer support, where it has gained no traction despite two all-hands presentations and executive endorsement. An organizational network analysis shows the three functions form largely separate clusters, connected by only four individuals, none of whom hold senior titles. What does network theory recommend?