What Is Social Network Analysis?
Social network analysis (SNA) is a research methodology that maps and measures relationships between people, teams, or organizations to understand how social structures influence behavior, information flow, and outcomes. SNA treats relationships as the primary unit of analysis rather than individual attributes, representing each person or entity as a node and each relationship (communication, advice-seeking, collaboration, trust) as an edge connecting them. The method originated in sociometry in the 1930s with Jacob Moreno's work on interpersonal connections and gained computational power in the 1990s as software made it feasible to analyze networks with thousands of nodes. Today, SNA is applied in organizational development, influence marketing, public health, community research, and anywhere the pattern of who connects to whom matters as much as what individuals think.
Why Social Network Analysis Matters
Organizational charts show formal reporting lines, but decisions, information, and influence travel through informal networks that may look nothing like the org chart. SNA reveals the actual communication structure, identifying who the real connectors and bottlenecks are. Rob Cross's research at the University of Virginia found that employees' positions in advice networks predicted their performance ratings more accurately than their job titles or tenure. In marketing, SNA helps identify opinion leaders whose adoption of a product influences their connections, amplifying word-of-mouth far beyond what individual targeting can achieve.
How Social Network Analysis Works
Collecting Relational Data
SNA requires data about who connects to whom and the nature of those connections. The most common collection method for organizational research is a network survey asking questions like "Who do you go to for work-related advice?" or "Who do you collaborate with at least weekly?" Respondents either name people freely (name generator) or select from a roster of all possible contacts. For marketing research, relational data might come from social media connections, referral chains, co-purchasing patterns, or survey questions about influence ("Whose recommendations do you trust for technology purchases?").
Mapping the Network
Once you have relational data, SNA software (Gephi, UCINET, NodeXL, or NetworkX in Python) creates a visual map. Nodes are sized by centrality, colored by attribute (department, customer segment, role), and positioned using algorithms that place connected nodes closer together. The resulting visualization immediately reveals clusters, isolates, bridges, and structural holes that aren't visible in tabular data.
Measuring Network Position
Several metrics quantify each actor's position. Degree centrality counts direct connections, identifying the most socially active nodes. Betweenness centrality identifies brokers who connect otherwise separate groups. Eigenvector centrality weights connections by the importance of the connected nodes, measuring not just how many connections someone has, but how well-connected their connections are. In marketing terms, a high-eigenvector influencer doesn't just have many followers; their followers have many followers too.
Identifying Structural Patterns
Beyond individual positions, SNA reveals structural patterns. Structural holes are gaps between clusters where no direct connections exist. The people who bridge structural holes control information flow between groups and often generate the most innovative ideas because they access diverse perspectives. Cliques are fully connected subgroups where everyone knows everyone, which promotes trust but can also create echo chambers. Core-periphery analysis identifies whether the network has a dense, active core with a loosely connected outer ring.
A Worked Example
A technology company with 400 employees across four offices ran an SNA survey asking about advice-seeking, innovation collaboration, and energy-giving relationships. The analysis found that the top 5% of employees by betweenness centrality handled 40% of cross-office information flow. Three employees in the engineering office were the sole bridges between R&D and the sales team. When one of them went on leave, cross-functional project velocity dropped measurably. The company used these findings to create formal liaison roles and cross-office working groups that reduced dependence on individual brokers.
When to Use Social Network Analysis
- Organizational network analysis mapping communication, collaboration, and influence patterns to identify informal leaders, bottlenecks, and isolated teams
- Influencer identification in marketing research, finding the individuals whose opinions and behaviors most strongly affect their connections
- Community health research mapping how health information or behaviors spread through social groups to design targeted interventions
- Change management understanding which informal networks need to be engaged for organizational changes to gain traction
- Knowledge management identifying where expertise sits and how it flows (or doesn't) across departments and geographies
Common Mistakes
- Surveying only part of the network produces incomplete maps where isolated-looking nodes may actually have many connections outside your sample boundary; aim for 80%+ response rates in bounded network studies
- Treating all relationships as equal when the strength and type of connection matters enormously; someone you email daily is a different kind of tie than someone you met once at a conference
- Presenting network maps without actionable recommendations because stakeholders find visualizations interesting but need specific interventions tied to the structural findings
How Quali-Fi Supports Social Network Analysis
Quali-Fi's survey builder supports roster-based and name-generator network questions that produce relational data ready for SNA tools. The platform handles complex question routing, allowing you to ask follow-up questions about each named connection (frequency, type of interaction, trust level) without overwhelming respondents.
Frequently Asked Questions
How large a network can I study with SNA?
Practical SNA studies range from small teams (15-30 people) to entire organizations (several thousand). For survey-based SNA, response burden increases with roster size, so organizations above 500 people typically use stratified approaches, surveying within and between departments rather than asking everyone about everyone. Digital trace data (email logs, Slack interactions) can handle much larger networks without survey burden.
What's the difference between SNA and standard survey analysis?
Standard surveys analyze individual attributes (each person's satisfaction, each person's engagement). SNA analyzes relationships between respondents. The unit of analysis shifts from the individual to the dyad (pair) or the network structure. This means you need different data collection methods, different metrics, and different statistical models.
Can SNA be done with anonymous surveys?
Not typically. SNA requires identifying who connects to whom, which means respondents must name (or select) specific people. Confidentiality protections apply to the analysis and reporting (individual positions aren't disclosed to management without consent), but the data collection itself can't be anonymous. This is a key ethical consideration that requires clear communication about how data will be used.
Related Topics
- Network Analysis
- Employee Engagement Data Analysis
- Cross-Tabulation Analysis
- Quantitative Content Analysis
- Data Collection Methods
Design network surveys that map real relationships -- try Quali-Fi free for 14 days.