Organizational networks: what the org chart can’t see
Who connects teams, where the silos are and what is lost when someone leaves.
The org chart says who reports to whom. Many hard decisions depend on something else: how work flows in practice. Knowing who you cannot afford to lose, which teams do not talk to each other or why a post-merger integration is stalling requires seeing that network, and that is what organizational network analysis (ONA) is for.
Org chart and network
The metrics and how to read them
Network science has dozens of metrics. For management, three are usually enough:
- Degree (how many ties a person has). It indicates collaboration overload. People with very high degree are often bottlenecks nobody sees, because almost everything runs through them, and they are at higher risk of burning out.
- Betweenness (how many paths run through a person). It indicates bridge risk. Whoever connects two teams that do not talk to each other holds knowledge that is not documented. If they resign, those teams are left disconnected.
- Between-group density (how much two teams talk to each other). It is used to measure silos. “Sales and Operations don’t coordinate” becomes a number with a baseline and a target.
Where the data comes from and the ethical care it requires
A network is built from two types of sources:
- Active: network surveys (“who do you go to in order to solve X?”, “who do you collaborate with every week?”). They take more fieldwork, but they come with explicit consent and give better data on trust and knowledge.
- Passive: collaboration metadata (calendars, email, work tools). Only who interacts with whom and how often is used, never the content.
Some rules must always hold: anonymization and aggregation by default, individual results only with consent, an informed ethics committee or employee representative, and a declared purpose before any data is collected. ONA is a tool for organizational design. If it is used for surveillance, people stop trusting it and later measurements lose validity.
What it is used for
- Restructurings and mergers. Before changing the org chart, it shows which bridges are about to break. Afterward, it lets you measure whether the teams are integrating in practice.
- Succession and retention. The people who are critical in the network are not always on the formal key-talent list. The analysis helps identify them before they resign.
- Onboarding. An early signal worth watching: new hires who have no ties beyond their own team by day 90. The network shows them in time to step in.
- Innovation and silos. It shows which teams lack bridges between them and, sometimes, where there are too many meetings.
What an ONA project looks like
It takes four to eight weeks. It includes defining the purpose and the ethical protocol, collection (a short survey and metadata where applicable), building and analyzing the network, and as the deliverable a list of decisions with an owner and a date: which bridges to protect, which silos to address and which overloads to redistribute. The network is measured again 6 to 12 months later to see how it changed.
The interactive ONA demo shows a collaboration network with bridges, silos and the “what happens if this person leaves?” simulation.
How the “what happens if this person leaves?” simulator works
Each dot is a person, sized by their number of ties. Three people hold 24.6% of the network’s brokerage.
Without T-04 no one is isolated, but the average distance between people rises 18.2%: from 3.95 to 4.67 steps.
Without A-03, 10 people in Administration & Finance are left with no path to the rest of the network. Dark red marks who is cut off.
One click puts the network back so you can try another profile.