Organizational network analysis — Organizational Development

Organizational network analysis (ONA) maps the informal structure: who actually goes to whom for information, advice, decisions, and trust. It reveals the organization as it works, rather than as it is drawn.

Evidence: Good as a method. Social network analysis is a mature, validated research methodology, and several of its organizational findings are robust: brokerage positions confer information advantage, weak ties carry novel information, and network position predicts performance and influence better than formal rank does. The weakness is practical rather than theoretical — response rates below roughly 70% make a map unreliable, and reported ties differ from actual ties.

Use when: the org chart does not explain observed behavior, when planning a change that needs informal sponsors, or when assessing what a reorganization or a wave of departures will actually break.

Do not use when: you cannot guarantee ethical handling. ONA data is unusually sensitive.

What is mapped

Different questions produce different, non-overlapping networks:

Information: who do you go to for the information you need to do your job?
Advice / expertise: whose technical judgment do you seek?
Trust: who would you tell about a mistake before it became public?
Energy: who leaves you more motivated after an interaction?
Innovation: who do you talk to when you have a half-formed idea?

The trust network is usually the smallest, the least like the org chart, and the most predictive of whether a change will take hold.

Key roles the map reveals

Central connectors: high in-degree; the people everyone consults. Often invisible to leadership, often overloaded, and often not senior.
Brokers / boundary spanners: connect otherwise separate clusters. Removing one disconnects parts of the organization. Rare and undervalued.
Peripheral nodes: weakly connected. Sometimes new joiners; sometimes people quietly disengaging; sometimes deep specialists who are fine.
Silos: dense internal ties, sparse external ties. Predicts handoff failures at the boundary. See conways-law.
Bottlenecks: single points of connection whose absence stops flow.

Methods

Survey-based: ask a roster question ("name up to five people you go to for X"). Explicit consent, best signal quality, snapshot in time.
Passive / digital exhaust: derive ties from email, chat, calendar, code review, or ticket data. Continuous and cheap, but ethically fraught and biased toward whatever tool you can read.

Prefer survey-based for OD work. Passive ONA without informed consent is surveillance, and will be experienced as such the moment it becomes known.

Ethics rules

Informed consent before collection, with a genuine opt-out.
Never report individual-level maps to managers about their reports.
Report roles and patterns, not named rankings, unless the named person has agreed.
Do not use ONA output in performance, promotion, or redundancy decisions. This is the single fastest way to destroy the data's validity and the practitioner's credibility.
Delete raw ties on a stated schedule.

See od-ethics.

Examples by audience

Health care: ONA reliably finds that a small number of senior nurses are the real information hubs of a ward, outranking the formal hierarchy. Change programs that recruit them succeed; programs that route only through consultants and managers stall.
Software: code review networks are a ready-made ONA dataset. A single reviewer approving most changes across many repositories is a bus-factor risk and a delivery constraint at the same time.
Executive: after an acquisition, ONA six months in shows whether the two organizations have actually merged. Two dense clusters with three thin links between them means integration exists only on paper.

Limitations

Snapshot; networks change quickly.
Response rate below roughly 70% makes the map unreliable.
Measures reported ties, which differ from actual ties.
Tempting to over-interpret pretty diagrams.

Questionnaire

A paired, copy-and-use diagnostic instrument for this model: organizational-network-analysis-questionnaire.