The Power of People Analytics: Identifying & Changing a Toxic Work Culture
People analytics refers to using data (from HR systems, surveys, communication tools, etc.) to generate insights about workforce behaviour, performance, and engagement.
It’s not just about tracking headcounts or performance ratings — advanced people analytics includes network analysis, natural language processing (NLP), and predictive modelling. Ethical design is also crucial: some researchers argue for “inverse transparency,” which would allow employees to see how their data is used to avoid unintended negative effects.
In this article, we’ll explore the power of people analytics and how in-house or consultant HR teams can use them to identify and change a toxic workplace.
Two Magpies is a trusted provider for workplace analytics and people-focused change. Find out more about how our services can help you in identifying and changing a toxic work culture in your organisation.
Quantitative Analytics for Identifying & Changing a Toxic Work Culture
“Hard metrics” — numbers that can be tracked and modelled — include:
Turnover and Attrition Metrics: High voluntary turnover in certain teams can flag areas of toxic culture. If people are leaving disproportionately from specific departments, this might indicate deeper issues.
Absence & Wellness Data: Elevated sickness absence, particularly related to stress or mental health, can be a strong quantitative indicator of cultural problems. CIPD research links poor culture to ill-health.
Engagement Survey Scores: Pulse surveys or annual engagement surveys are classic tools. Analytics teams can track trends in engagement over time, segment by team/manager, and identify “hotspots” where scores are persistently low or deteriorating.
Network / Communication Analytics: Using tools like Organisational Network Analysis (ONA) (for example, via platforms like Humanise) to map how people actually communicate. Patterns such as siloing, exclusion, or overly centralised communication flows may signal cultural problems.
Predictive Modelling: Advanced people analytics teams (especially mature ones) build predictive models to forecast risk (e.g., attrition, burnout) using multiple data sources. McKinsey describes how top teams “make reliable, consistent, and valid predictive analytics.”
Data-Centric Culture Correlation: According to Deloitte, having a “data-centric culture” is strongly associated with people analytics maturity and better outcomes.
Questionnaire-Based Culture Profiling: There is validated research that uses structured, quantitative questionnaires to assess organisational culture. For example, Marchand et al. used a validated “Organisational Culture Profile” to measure how different types of culture correlate with psychological distress, exhaustion, etc.
Using Qualitative Analytics (People Data) to Detect Subtle Toxicity
Quantitative data gives signals, but qualitative analytics (free-text, conversations, interviews) can uncover why there might be toxicity.
Text Analysis of Employee Feedback: Use NLP on open-ended survey responses, exit interview transcripts, or internal feedback to spot recurring negative themes (e.g., “blame,” “micromanagement,” “fear,” “no trust”).
Conversation / Voice Analytics: Tools like Insight7 analyse real employee-manager conversations, chats, or voice communications to detect patterns of toxicity in language, sentiment, or interaction style.
Interviews & Focus Groups: Qualitative research through interviews or focus groups helps dig into the intangible aspects of culture (power dynamics, psychological safety, underlying norms). For instance, an interpretive review in healthcare found that many “organisational culture” tools fail to capture deep “intangible” themes like trust, blame, cohesion, and psychological safety.
Sentiment & Theme Tracking Over Time: By combining regular pulse surveys and free-text questions, analytics teams can track how sentiments evolve, whether areas improve or worsen, and correlate them with interventions (or lack thereof).
Why Analytics Is Powerful for Highlighting Toxic Culture
People analytics is powerful for identifying and changing a toxic work culture. Here’s how.
Early Detection and Proactive Intervention
People analytics allows organisations to spot risk before culture issues become deeply embedded. For example, rising absence + negative sentiment = early red flags.
Predictive models may identify teams at risk of burnout or high turnover, enabling leaders to intervene early.
Evidence-Based Diagnosis
Rather than relying on anecdotes (e.g., “this manager is bad”), analytics provides data to back up claims. This gives HR or leadership credibility when raising cultural issues.
Qualitative data helps validate and explain what’s going on; it’s not just numbers but lived experience.
Targeted and Tailored Action
By identifying which teams or subcultures are most at risk, organisations can allocate resources (coaching, mediation, management training) more effectively.
Analytics helps monitor the impact of interventions: if you run a training programme, you can track sentiment and turnover before and after the intervention.
Transparency & Accountability
When analytics is done responsibly (with transparency), it helps build trust: employees can see how their feedback is used, and leaders can be held accountable for remedying issues.
Analytics can also show whether leadership interventions are working: are sentiment and behaviour improving over time?
Scalability
Unlike one-off surveys or sporadic HR investigations, analytics can scale across the organisation. Once set up, dashboards can run continuously, tracking culture in real time (or near real time).
Risks, Challenges & Ethical Considerations When Changing a Toxic Work Culture
Privacy and surveillance concerns arise because people may feel “watched” if their communication is being analysed, which means analytics teams must maintain robust governance, apply strong anonymisation practices, and communicate transparently about how data is used.
Bias in data and models is another risk, as predictive systems can unintentionally reinforce existing inequities or mislabel individuals as “at risk” when safeguards and regular checks are not in place.
There is also a danger of over-relying on quantitative signals, since metrics alone cannot capture the full context. Without qualitative insight, organisations may end up “solving for the number” while missing the underlying causes of problems.
Data quality issues further complicate people analytics work; many organisations struggle with incomplete or inconsistent HR data, which weakens the reliability of any insights and highlights the need for strong data foundations.
Resistance from leaders can emerge when analytics surfaces evidence of a toxic culture. Without a foundation of trust and transparency, these insights may be ignored or rejected rather than translated into meaningful action.
Real-World & Emerging Examples
People analytics firms such as Humu use survey data and behavioural “nudges” to encourage managers to hold meaningful conversations after survey results are released, helping ensure that insights are actually translated into action.
These platforms can also analyse employee communications — including calls and chat messages — to identify toxicity patterns by examining language use and sentiment.
Organisations that use human-centric organisational network analysis tools, such as those offered by Humanise, map collaboration and communication patterns to understand how people interact; when the analysis reveals outlier, isolated, or tightly clustered networks, it may indicate exclusion, power imbalances, or the presence of distinct microcultures.
In the UK, Oak Engage’s “Toxic Workplace Report” drew on survey data to demonstrate that 75% of workers had experienced some form of toxic culture, providing a large-scale quantitative signal about the prevalence of such issues.
Recommendations for Using People Analytics to Detect Toxic Culture
Build a Cross-Functional Analytics Team
Combine HR, data science, and organisational psychology expertise to interpret both quantitative and qualitative data.
Follow McKinsey’s model: use data engineers, data scientists (particularly NLP or network analytics), and domain experts.
Establish Ethical Governance
Create a people data governance framework: explicit consent, anonymisation, transparent usage.
Consider “inverse transparency” — letting employees see what data is collected, how it’s used, and what insights are drawn.
Collect Mixed Data
Use engagement/pulse surveys with both quantitative (e.g., Likert scales) and open-text items.
Capture communication data (where legally and ethically possible): internal chats, email metadata, network data, meeting patterns.
Run focus groups or interviews periodically to get a deeper context.
Build Diagnostics and Dashboards
Develop culture-risk dashboards (e.g., attrition risk, sentiment trends, network isolation).
Set up alerting for red flags: rising absence, low psychological safety, and negative sentiment themes recurring.
Act on Insights
Use insights to target interventions: coaching, mediation, leadership development, and workload adjustments.
Involve leaders in interpreting data and setting action plans — not just analytics teams.
Monitor interventions: track cultural metrics before and after, to measure impact.
Communicate Transparently
Share high-level insights with the workforce (anonymised): what you’ve found, what you plan to do, and why.
Foster a feedback loop: employees should be able to react to the data, suggest actions, and see progress.
Contact Two Magpies to learn more about our workplace analytics and mediation services.