People Analytics and Trust: When Transparency Reveals Too Much
- Jonathan H. Westover, PhD
- 8 hours ago
- 20 min read
Listen to a review of this article:
Abstract: Organizations increasingly deploy commercial people analytics (PA) systems to inform workforce decisions, yet fundamental questions remain about how these systems shape employee–employer relationships. This study examines how awareness of information asymmetries created by PA influences employee trust and retention intentions. Using a scenario-based experiment with German knowledge workers (N = 438), we find that PA adoption significantly erodes organizational trust and increases turnover intentions—effects driven primarily by privacy concerns rather than system sophistication. Employees exposed to the full scope of managerial dashboards (Study 1) report substantially worse perceptions than those seeing only employee-facing interfaces (Study 2), revealing how transparency about algorithmic monitoring paradoxically undermines trust. These findings challenge vendor claims that PA enhances employee wellbeing and suggest that current implementations reverse traditional information asymmetries in ways employees find deeply troubling, even when they cannot opt out.
When Microsoft announced in 2024 that its Viva Insights platform had reached 66 million users—just three years after launch—it marked a watershed moment for workplace analytics (Foss, 2024). People analytics systems, which collect and analyze employee behavior data to inform management decisions, have moved from experimental novelty to mainstream practice with remarkable speed. Recent OECD research indicates that algorithmic management tools are now far more prevalent than previously estimated across developed economies (Milanez et al., 2025), touching millions of knowledge workers' daily experience.
The promise is compelling: data-driven insights to optimize performance, reduce bias, and support employee wellbeing. The reality is more complex. These systems create a fundamental shift in workplace power dynamics by reversing long-standing information asymmetries. Traditionally, employees possessed superior knowledge about their own efforts, work processes, and productivity—a cornerstone assumption of principal-agent theory (Jensen & Meckling, 1976). PA systems upend this arrangement, providing managers with granular, algorithmic insights into employee behavior that workers themselves cannot access or contest.
This reversal carries profound implications that remain underexplored. While researchers have documented risks ranging from algorithmic bias to privacy violations (Giermindl et al., 2022; Kordzadeh & Ghasemaghaei, 2022), we know surprisingly little about how employees perceive organizations that deploy these systems—especially when workers become aware of the disparity between what they see and what management knows. Do employees accept this surveillance as a reasonable extension of managerial prerogative? Or does awareness of algorithmic monitoring erode the trust essential for productive employment relationships?
These questions matter because the discourse surrounding PA often emphasizes transparency as a solution to employee concerns (Ngwenyama et al., 2024). Yet transparency alone may prove insufficient or even counterproductive if it reveals information asymmetries that employees find fundamentally unfair but cannot refuse. Unlike consumers who can decline to use a particular app or platform, employees face limited options: accept the monitoring or leave the organization.
This study investigates how PA adoption influences employee trust and retention intentions, with particular attention to the role of information asymmetry awareness. We examine whether employees respond differently when confronted with the full scope of managerial analytics versus only employee-facing interfaces. Additionally, we test whether system sophistication—ranging from descriptive statistics to predictive recommendations—affects employee perceptions, given that regulations and ethics guidelines often distinguish between these analytical levels (European Commission, 2021).
Understanding these dynamics is increasingly urgent as PA becomes standard practice and as organizations navigate the tension between data-driven decision-making and employee expectations of privacy, fairness, and respect.
The People Analytics Landscape
Defining People Analytics in Contemporary Organizations
People analytics encompasses a diverse array of practices, making precise definition challenging. We adopt McCartney and Fu's (2022) conceptualization: PA represents an ongoing organizational process of collecting, transforming, and deploying workforce data to generate insights, performed at varying analytical sophistication levels. This process manifests through commercial software systems that automate data collection and analysis, providing managers with dashboards of employee metrics.
Leading vendors like Microsoft and SAP dominate the market, with their solutions embedded in enterprise software millions of workers use daily. Microsoft Viva exemplifies the dual-interface design common to many PA systems: employees receive dashboards featuring wellness reminders, mood tracking, and productivity suggestions, while managers access performance analytics, retention risk scores, and team collaboration metrics. This architectural separation creates the information asymmetry central to our investigation.
The data feeding these systems extend far beyond traditional HR metrics. Modern PA captures "digital exhaust"—the behavioral traces workers leave through email patterns, meeting attendance, calendar usage, instant messaging, and even keystroke dynamics (Leonardi, 2021). Time-stamped interactions, communication networks, and work-hour patterns generate rich longitudinal datasets that algorithms analyze to identify productivity patterns, burnout risk, flight risk, and collaboration bottlenecks.
Common PA use cases span the employee lifecycle: recruitment screening, onboarding effectiveness, performance evaluation, development recommendations, retention risk assessment, and organizational network analysis (Gal et al., 2020; Tursunbayeva et al., 2018). The systems promise to make visible what was previously opaque—who contributes most to projects, which managers develop talent effectively, where collaboration breaks down, and why employees leave.
Prevalence, Drivers, and Adoption Patterns
PA adoption has accelerated dramatically, driven by several convergent trends. The COVID-19 pandemic normalized remote work arrangements that, paradoxically, intensified demand for monitoring tools to maintain visibility into distributed teams (Aloisi & De Stefano, 2022). Organizations already collecting workforce data through enterprise systems found incremental costs of PA adoption relatively low, while vendors aggressively marketed PA as essential for modern management.
The OECD's 2025 employer survey revealed that PA adoption substantially exceeds prior estimates in both Europe and North America, with particularly high prevalence in sectors like professional services, technology, and finance (Milanez et al., 2025). However, adoption remains uneven, with many organizations still deliberating implementation or piloting limited functionality. This transitional state creates a natural experiment: workers can currently compare employers with and without PA systems, though this comparison may fade as adoption reaches saturation.
Germany, our study context, presents an instructive case. Strong co-determination traditions and data protection frameworks create institutional constraints on PA implementation, yet adoption proceeds nonetheless. Works councils possess statutory rights to negotiate PA deployment and access external expertise (Bundesministerium für Arbeit und Soziales, 2021), though this protection applies only to workplaces with established worker representation—a minority of German firms.
Adoption drivers reflect both genuine organizational needs and vendor-driven demand creation. Organizations seek productivity insights, retention predictors, and efficiency gains in genuine pursuit of performance improvement. Yet vendors also promote PA through compelling narratives about "data-driven HR," "predictive workforce planning," and "employee experience optimization" that may oversell actual capabilities while understating risks (Ngwenyama et al., 2024).
Organizational and Individual Consequences of People Analytics
Organizational Performance Claims and Realities
PA vendors promise substantial organizational benefits: improved hiring quality, reduced turnover, enhanced productivity, and better workforce planning. Some evidence supports these claims. Organizations using analytics-informed retention strategies report modestly lower turnover in certain employee segments, while firms applying structured data analysis to recruitment sometimes achieve better hiring outcomes than those relying solely on intuition (Tursunbayeva et al., 2018).
However, the evidence base remains thin relative to vendor promises. Many claimed benefits lack rigorous evaluation, instead relying on vendor case studies, consultant reports, or correlational analyses that cannot establish causality. The "black box" nature of many PA algorithms makes it difficult even for deploying organizations to verify whether promised outcomes materialize or whether observed improvements stem from PA rather than confounding factors.
Moreover, documented risks suggest PA may harm organizational performance through multiple pathways. If PA adoption erodes employee trust and increases turnover intentions—as our findings suggest—productivity gains from optimization may be offset by engagement losses, knowledge attrition, and recruitment costs. Organizations may also experience reduced innovation and collaboration if employees perceive constant monitoring and adjust behavior accordingly, a dynamic Bernstein (2017) documented in manufacturing contexts.
The integration of PA into managerial practice remains immature. Managers often lack training in interpreting analytics, may overweight algorithmic recommendations relative to contextual knowledge, or may deploy PA insights inequitably (Ruschemeier & Hondrich, 2024). These implementation challenges can convert potentially useful data into sources of bias, unfairness, and demotivation.
Employee Wellbeing and Experience Impacts
Vendors market PA as supporting employee wellbeing through personalized recommendations, burnout detection, and workload balancing. Employee-facing dashboards feature wellness prompts, focus-time suggestions, and mood-tracking tools presented as individual empowerment resources. However, research reveals concerning effects on worker experience.
Constant monitoring creates psychological strain even when workers believe themselves observed for beneficent purposes. The awareness that algorithms continuously evaluate one's performance, relationships, and potential creates what Zuboff (2015) termed "anticipatory conformity"—workers preemptively adjust behavior to satisfy perceived algorithmic preferences, reducing authenticity and autonomy. Parent-Rocheleau and Parker (2021) found that algorithmic management correlates with increased stress and reduced job satisfaction across diverse contexts.
PA can exacerbate power imbalances and perceptions of unfairness. When managers access granular performance data unavailable to workers, employees may struggle to understand or contest evaluations. The metrics PA surfaces—emails sent, meetings attended, after-hours activity—capture only proxies for valuable work, yet their quantification imbues them with false precision (Rahman, 2021). Employees performing unmeasured but important work (mentoring, creative thinking, relationship building) become invisible to PA systems, potentially disadvantaging them in evaluations.
Privacy concerns represent another significant stressor. Research consistently shows employees harbor substantial worries about how organizations collect, analyze, and use their data (Tursunbayeva et al., 2022). These concerns intensify when employees lack transparency about data practices or when they cannot access the same insights about themselves that managers receive. The resulting privacy calculus—weighing surveillance costs against promised benefits—tends to skew negative for many workers (Bhave et al., 2020).
For employees who find PA particularly objectionable, options remain limited. Workplace monitoring differs fundamentally from consumer contexts where individuals might refuse to use privacy-invasive apps. Employees cannot unilaterally opt out of PA without jeopardizing their employment. This forced participation may generate resentment that undermines the employment relationship, even when employees never articulate their concerns openly.
The consequences of sustained intention to leave while remaining organizationally embedded are particularly troubling. Allen and colleagues (2016) documented that workers who wish to leave but cannot exhibit concerning behavioral and health patterns: increased absenteeism, reduced performance, emotional exhaustion, sleep problems, and self-destructive behaviors including elevated alcohol consumption. PA systems that create turnover intentions without enabling actual departure may thus harm worker wellbeing substantially.
Evidence-Based Organizational Responses
Table 1: People Analytics Implementation Strategies and Case Studies
Organization | PA Implementation Strategy | Governance Mechanism | Employee Data Access Level | Outcome or Reported Benefit | Key Challenges Identified |
Henkel | Bidirectional transparency providing team leaders efficiency metrics while giving individuals access to detailed performance data. | Works council approval and consultation. | Shared access: employees view the same metrics as managers and compare against anonymized team distributions. | Secured works council approval; generated more positive employee reception than initially anticipated. | Initial concerns over data accuracy and managerial surveillance. |
Schneider Electric | Establishment of a multinational governance body to vet all PA applications prior to deployment. | Digital Rights Committee (includes worker representatives, data protection officers, and business leaders). | Employees can petition the committee to review specific PA uses they find problematic. | Substantially higher employee trust in the systems that are deployed. | Slower PA adoption due to rigorous vetting processes. |
Siemens AG | Deployment governed by global "People Analytics Principles," including annual independent bias audits. | Works council approval and "Right to Explanation" for employees. | Employees can request detailed accounts of how PA influenced decisions affecting them. | Structured oversight and constraints on misuse; prohibition of PA use for employee termination. | Managing deployment across global operations with varying labor regulations. |
Microsoft | Quarterly internal impact reviews of its own Viva platform to assess sentiment and unintended consequences. | Viva Impact Reviews (assessing accuracy and sentiment). | Expanded employee data access following reviews; retirement of problematic analytics. | Feature modifications and the retirement of problematic analytics based on employee feedback. | Identification of unintended consequences and problematic analytics within the suite. |
Volkswagen | Partnering with unions for joint technical training on algorithmic management systems. | Joint training with IG Metall union to build shared analytical capability. | Technical training for worker representatives to enable independent evaluation of system data. | Facilitated more productive negotiations regarding PA boundaries and applications. | Information asymmetry between management and labor representatives. |
Accenture | Mandatory algorithmic literacy and ethics training for all managers accessing PA systems. | Assessment-based access control (three-module training required for detailed data access). | Individual data access is restricted to managers who pass assessment; others see only aggregated team metrics. | More nuanced PA-informed decisions and fewer employee complaints about unfair algorithmic treatment. | Risk of uninformed or biased human interpretation of analytics. |
Mastercard | Collaboration analytics to improve meeting efficiency utilizing a "data dividend" model. | Benefit sharing: converting estimated time saved into tangible employee rewards. | Not in source | Reframed PA from a surveillance tool to a collective benefit; improved employee reception. | Overcoming the perception of PA as purely a surveillance tool. |
Organizations implementing or considering PA adoption face complex choices about system design, deployment approach, and governance structures. Evidence from research and practice suggests several strategies for addressing the trust and fairness concerns that PA raises.
Transparent Communication and Shared Access
Research foundations: Information asymmetry emerges as a critical mechanism driving negative PA perceptions. When employees understand that managers access substantially different and more detailed information than they do, trust erodes (Klöpper & Rowe, 2024). Conversely, genuine transparency—not merely disclosure of data collection but meaningful explanation of analytical processes and shared access to insights—can mitigate concerns.
Practical approaches include:
Employee analytics dashboards with parity: Provide employees with access to the same performance metrics, relationship analytics, and predictive insights that their managers receive, not merely gamified wellness prompts
Clear, accessible data dictionaries: Explain precisely what data the system collects, how algorithms process information, which metrics drive which decisions, and where analytical boundaries exist
Algorithmic explainability interfaces: Where PA generates predictions or recommendations, provide explanations that employees can understand and interrogate regarding what factors influenced algorithmic judgments
Regular transparency reports: Organizations can issue periodic reports detailing PA system usage, what decisions relied on analytics, how accuracy was validated, and what errors or biases were detected and corrected
Henkel exemplifies shared-access approaches in its German operations. The chemical manufacturer's workforce analytics initiative provides team leaders with collaboration and efficiency metrics, but also grants individual contributors access to their own detailed performance data and the algorithmic factors influencing their ratings. Employees can compare their metrics against anonymized team distributions and request explanations for ratings they consider inaccurate. This bidirectional transparency helped secure works council approval and generated more positive employee reception than initially anticipated.
Procedural Justice and Meaningful Consent
Research foundations: Procedural justice research demonstrates that people often accept unfavorable outcomes when they perceive decision processes as fair (Tong et al., 2021). For PA, procedural justice requires more than legal compliance or pro forma consent; it demands genuine employee voice in system design and deployment.
Practical approaches include:
Participatory design processes: Involve employees and worker representatives in decisions about which data to collect, what analytics to deploy, and how to present and use insights
Opt-in pilot programs: Where feasible, begin PA deployment with voluntary participation, gathering feedback and demonstrating value before mandating universal adoption
Ethics review boards with worker representation: Establish standing committees including employee representatives to review new PA applications, monitor for unfair impacts, and adjudicate concerns
Appeal and correction mechanisms: Create clear processes through which employees can challenge PA-based decisions, request human review, or correct data they believe inaccurate
Schneider Electric, the French energy management multinational, established a "Digital Rights Committee" when deploying PA across its European operations. The committee includes worker representatives, data protection officers, and business leaders, and must approve any new PA application before deployment. Employees can also petition the committee to review specific PA uses they find problematic. This structure has generated both slower PA adoption and substantially higher employee trust in the systems that do deploy.
Algorithmic Capability-Building for Managers and Employees
Research foundations: Much PA harm stems not from the technology itself but from uninformed or biased human interpretation and application (Ghasemaghaei & Kordzadeh, 2024). Managers may overweight algorithmic recommendations, misinterpret statistical outputs, or fail to recognize bias. Employees lack literacy to evaluate algorithmic assessments critically.
Practical approaches include:
Manager training on algorithm limitations: Educate leaders about what PA systems can and cannot reliably assess, how algorithmic bias emerges, why context matters in interpretation, and when to override recommendations
Data literacy programs for all employees: Provide workforce training in understanding statistics, interpreting visualizations, recognizing spurious correlations, and questioning algorithmic authority
Decision-making protocols that combine human and algorithmic judgment: Establish organizational norms that PA insights inform rather than determine decisions, with explicit documentation of human reasoning
Regular algorithmic audits: Periodically evaluate whether PA systems produce biased, inaccurate, or unfair outputs, with results shared transparently across the organization
Accenture, the professional services firm, requires all managers with access to PA systems to complete a three-module training program covering algorithmic literacy, bias detection, and ethical application. Managers who fail assessment cannot access detailed workforce analytics, receiving only aggregated team metrics instead. The company reports that trained managers make more nuanced PA-informed decisions and generate fewer employee complaints about unfair algorithmic treatment.
Operating Model Adjustments and Algorithmic Governance
Research foundations: PA systems exist within broader organizational contexts of power, culture, and governance. Their impacts depend critically on how organizations integrate analytics into decision-making structures and whether effective oversight constrains misuse (Jarrahi et al., 2021).
Practical approaches include:
Algorithmic impact assessments: Before deploying PA applications, conduct structured evaluations of potential impacts on different employee groups, with particular attention to discrimination risks
Limited data retention: Establish and enforce policies restricting how long the organization retains granular employee behavior data, with automatic deletion after reasonable periods
Purpose limitation and scope boundaries: Clearly define and communicate which decisions PA may inform (e.g., workforce planning) versus which remain off-limits (e.g., individual termination decisions)
Third-party auditing: Engage external auditors to review PA systems periodically for bias, accuracy, and compliance with organizational policies and legal requirements
Siemens AG, the German industrial conglomerate, developed comprehensive "People Analytics Principles" governing PA deployment across its global operations. These principles prohibit using PA for individual termination decisions, mandate annual bias audits of all PA systems by independent experts, and require works council approval for new analytics applications in Germany (with equivalent consultation elsewhere). The principles also establish a "right to explanation" through which employees can request detailed accounts of how PA influenced decisions affecting them.
Compensation and Benefit Sharing
Research foundations: When organizations profit from employee data, questions of fair distribution arise. Li and colleagues (2023) argue for reconsidering data ownership in contexts where firms extract value from individual information. If PA systems genuinely enhance productivity, workers who generate the data enabling these gains deserve to share benefits.
Practical approaches include:
Productivity-linked compensation adjustments: Where PA demonstrably improves organizational performance, share gains through enhanced compensation, bonuses, or profit-sharing
Reduced working hours: If PA enables efficiency gains, consider reducing required working hours while maintaining compensation—a practical application of four-day workweek concepts
Enhanced development resources: Allocate PA-driven savings to employee development programs, improving career prospects for those whose data feed analytics
Data rights and compensation: In forward-looking approaches, consider compensating employees directly for valuable data contributions, similar to research participant payments
Mastercard implemented a novel approach when deploying collaboration analytics to improve meeting efficiency. The company tracked estimated time saved through PA-informed meeting reduction and converted these savings into additional paid time off distributed to participating employees. This "data dividend" approach helped reframe PA from surveillance tool to collective benefit, substantially improving employee reception.
Building Long-Term Organizational Capabilities
Beyond tactical responses to immediate PA concerns, organizations should consider how to build enduring capabilities for navigating the ongoing evolution of workplace algorithms. Three strategic pillars emerge from research and practice.
Trust Recalibration and Psychological Contract Evolution
Challenge: PA fundamentally changes the employer–employee psychological contract—the unwritten expectations and obligations that govern workplace relationships (Rousseau, 1989). Employees may experience PA adoption as a breach of implicit commitments about privacy, autonomy, and respect. Organizations must actively renegotiate these contracts rather than assuming acceptance.
Strategic approaches:
Explicit psychological contract discussions: Create structured opportunities for employees to voice expectations about PA use, with organizational responses that either accommodate concerns or explain why certain practices remain necessary
Values-based PA governance: Ground PA policies in articulated organizational values (e.g., respect, fairness, empowerment) and regularly evaluate whether practices align with stated commitments
Trust-building through consistent action: Recognize that trust erodes quickly but rebuilds slowly; organizations must demonstrate through sustained behavior that they use PA responsibly, not merely claim they will
Acknowledgment of legitimate concerns: Rather than dismissing employee worries about PA as technophobia or misunderstanding, validate that monitoring creates genuine tensions worthy of ongoing attention and adjustment
Organizations in high-trust cultures, particularly those with strong traditions of employee participation and worker representation, may find PA implementation less fraught if approached transparently. Research in software companies with deeply collaborative cultures found that bidirectional PA transparency—where employees and managers share equal access to workforce analytics—can actually enhance trust by making implicit patterns explicit and actionable (Gierlich-Joas et al., 2024). However, such outcomes require preexisting trust and participatory cultures; PA alone does not create these conditions.
Distributed Leadership and Power-Sharing Structures
Challenge: PA concentrates power in management by creating information asymmetry that reverses traditional principal-agent dynamics. Counterbalancing this concentration requires deliberately distributing algorithmic access and influence across organizational levels.
Strategic approaches:
Worker analytics councils: Establish employee-led committees with access to aggregated PA insights, empowered to conduct independent analyses, commission studies, and present findings to leadership
Union and works council PA access: Provide worker representatives with technical training and system access enabling independent evaluation of whether PA deployment aligns with negotiated agreements
Cross-functional algorithm review: Include employees from multiple levels and functions in governance bodies that approve new PA applications, evaluate ongoing systems, and mandate changes when problems emerge
Distributed expertise development: Invest in building PA literacy throughout the organization rather than concentrating knowledge in management and HR, enabling informed dialogue about appropriate uses
IG Metall, Germany's metalworkers union, has pioneered "algorithmic literacy" training for works council representatives, teaching them to access, interpret, and challenge PA systems deployed in organized workplaces. Some progressive employers like Volkswagen have partnered with the union to provide joint training that builds shared vocabulary and analytical capability, facilitating more productive negotiations about PA boundaries and applications.
The emerging concept of "data-driven labor organizing" extends this logic further (Nyman et al., 2024). When unions or workers councils possess analytical capability comparable to management's, they can leverage the same workforce data that PA systems collect to identify unsafe conditions, document pay inequities, prove discrimination, and advocate for improved working conditions. This approach treats PA as a potentially democratizing technology if access and expertise are distributed equitably.
Adaptive Learning and Continuous Recalibration
Challenge: PA technology evolves rapidly, with new analytical capabilities, data sources, and applications emerging constantly. Static policies and one-time governance decisions cannot keep pace. Organizations require dynamic systems for continuously evaluating and adjusting PA practices.
Strategic approaches:
Regular stakeholder feedback loops: Institutionalize periodic surveys, focus groups, and consultation processes through which employees assess PA impacts and suggest modifications
Algorithmic agility with stability: Balance the need to adapt PA systems in response to feedback with the stability employees require to understand and trust consistent practices
Ongoing ethics scanning: Monitor research literature, regulatory developments, and peer organization experiences to identify emerging PA risks and promising governance approaches
Sunset provisions and renewal requirements: Implement automatic expiration dates for PA applications, requiring explicit renewal decisions that force periodic evaluation of whether continuing specific analytics remains justified
Microsoft, despite being a major PA vendor, has established relatively mature internal governance for its own PA deployment. The company conducts quarterly "Viva Impact Reviews" assessing employee sentiment toward the Viva suite, algorithmic accuracy metrics, and identification of unintended consequences. These reviews have led to multiple feature modifications, expanded employee data access, and retired analytics that proved problematic—demonstrating that even enthusiastic PA adopters benefit from structured reflection and adjustment (Boyd, 2025).
Conclusion
People analytics systems promise to illuminate the "black box" of organizational performance through data-driven insights into employee behavior, collaboration, and productivity. Yet our research reveals a profound irony: these transparency tools can undermine the trust essential for the very performance they aim to optimize. When employees become aware that PA creates information asymmetries favoring management—providing leaders with granular behavioral insights employees themselves cannot access—privacy concerns intensify, organizational trust erodes, and turnover intentions rise.
These effects prove surprisingly insensitive to system sophistication. Employees perceive descriptive analytics (basic statistics) as negatively as predictive or prescriptive systems (machine learning recommendations), suggesting that concerns center not on algorithmic complexity but on the fundamental power imbalance PA creates. Simply knowing that managers see more than they do troubles employees sufficiently to damage organizational relationships.
The comparison between employees exposed to full PA transparency and those seeing only employee-facing interfaces reveals the double bind organizations face. When employees remain unaware of management dashboards' scope, perceptions remain more positive—but this "ignorance is bliss" finding offers cold comfort. It implies that maintaining employee acceptance requires information asymmetry not just about their behavior but about PA systems themselves, a troubling foundation for trust-based employment relationships.
These findings carry significant implications for the millions of workers now subject to PA monitoring. The intention to leave that PA generates—even when workers cannot actually depart—creates documented harms to employee wellbeing, performance, and health (Allen et al., 2016). Organizations deploying PA may inadvertently inflict these costs on workers while believing they pursue data-driven management excellence.
For organizations, our results suggest PA systems in their current form present risks that may outweigh benefits. If PA adoption erodes trust sufficiently to increase turnover and reduce engagement, productivity gains from optimization may prove pyrrhic. Organizations should approach PA deployment with considerable caution, investing heavily in transparency, procedural justice, shared access, and governance structures that constrain misuse.
The ultimate question remains whether PA can be reimagined to serve employee interests rather than merely providing management with enhanced control. Bidirectional transparency—where workers access the same insights as their leaders—represents one promising direction, particularly in high-trust organizational cultures (Gierlich-Joas et al., 2024). Data-driven labor organizing, where unions leverage PA infrastructure to advocate for workers, offers another (Nyman et al., 2024). Both approaches require fundamental shifts in how organizations conceptualize PA's purpose and design its implementation.
Policymakers and regulators face parallel challenges. Current approaches that distinguish PA systems by technical sophistication (descriptive vs. predictive vs. prescriptive) may miss the mark if all levels equally concern employees. More promising regulatory frameworks might focus on information symmetry requirements, mandating that workers receive substantially equivalent analytics access as their managers, or establish strong consent regimes that enable genuine refusal of monitoring without employment consequences.
The vendor discourse promoting PA as transparent, objective, and employee-centric requires critical examination and regulatory scrutiny. When marketing claims systematically understate surveillance implications while overstating wellbeing benefits, they prevent the informed evaluation essential for functioning labor markets. Regulation of PA marketing—similar to greenwashing rules for environmental claims—could serve public interest by enabling more honest assessment of PA's costs and benefits.
As PA adoption accelerates toward ubiquity, the window for shaping its trajectory narrows. The current moment, where workers can still compare monitored and unmonitored workplaces, offers crucial opportunities for learning about PA impacts and for establishing governance frameworks before fait accompli. Our research suggests that absent significant changes in PA system design, deployment practices, and power dynamics, widespread adoption risks normalizing workplace surveillance that employees find fundamentally troubling but cannot escape—a concerning prospect for the future of knowledge work.
The path forward requires moving beyond technical solutions to engage seriously with PA as a socio-technical phenomenon embedded in workplace power relations. Organizations that approach PA with humility, genuine transparency, distributed governance, and willingness to forgo analytics that prove harmful may discover ways to realize PA's potential benefits without destroying the trust that enables effective organizations. Those that deploy PA as merely enhanced surveillance will likely discover that datafication cannot substitute for the human relationships that remain essential to organizational success.
Research Infographic
References
Ajunwa, I., Crawford, K., & Schultz, J. (2017). Limitless worker surveillance. California Law Review, 105(1), 101–142.
Allen, D. G., Peltokorpi, V., & Rubenstein, A. L. (2016). When "embedded" means "stuck": Moderating effects of job embeddedness in adverse work environments. Journal of Applied Psychology, 101(12), 1670–1686.
Aloisi, A., & De Stefano, V. (2022). Essential jobs, remote work and digital surveillance: Addressing the COVID-19 pandemic panopticon. International Labour Review, 161(2), 289–314.
Bernstein, E. S. (2017). Making transparency transparent: The evolution of observation in management theory. Academy of Management Annals, 11(1), 217–266.
Bhave, D. P., Teo, L. H., & Dalal, R. S. (2020). Privacy at work: A review and a research agenda for a contested terrain. Journal of Management, 46(1), 127–164.
Boyd, K. (2025, January 15). Fostering employee wellbeing and improving productivity at Microsoft with Microsoft Viva Insights. Microsoft Inside Track. https://www.microsoft.com/insidetrack/blog/fostering-employee-wellbeing-and-improving-productivity-at-microsoft-with-microsoft-viva-insights/
Bundesministerium für Arbeit und Soziales. (2021). Betriebsrätemodernisierungsgesetz. German Federal Ministry of Labour and Social Affairs.
European Commission. (2021). Proposal for a regulation laying down harmonised rules on artificial intelligence. European Commission.
Foss, J. (2024, November 19). Ignite 2024: New innovation for the AI-powered workforce. Microsoft Viva Blog. https://techcommunity.microsoft.com/blog/microsoftvivablog/ignite-2024-new-innovation-for-the-ai-powered-workforce/4303391
Gal, U., Jensen, T. B., & Stein, M. K. (2020). Breaking the vicious cycle of algorithmic management: A virtue ethics approach to people analytics. Information and Organization, 30(2), 100301.
Ghasemaghaei, M., & Kordzadeh, N. (2024). Ethics in the age of algorithms: Unravelling the impact of algorithmic unfairness on data analytics recommendation acceptance. Information Systems Journal, online first.
Gierlich-Joas, M., Baiyere, A., & Hess, T. (2024). Inverse transparency and the quest for empowerment through the design of digital workplace technologies. Journal of the Association for Information Systems, 25(5), 1212–1239.
Giermindl, L. M., Strich, F., Christ, O., Leicht-Deobald, U., & Redzepi, A. (2022). The dark sides of people analytics: Reviewing the perils for organisations and employees. European Journal of Information Systems, 31(3), 410–435.
Jarrahi, M. H., Newlands, G., Lee, M. K., Wolf, C. T., Kinder, E., & Sutherland, W. (2021). Algorithmic management in a work context. Big Data & Society, 8(2), 1–24.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305–360.
Klöpper, M., & Rowe, F. (2024). People analytics, trust erosion, and intention to leave: The role of information asymmetry. Information & Management, forthcoming.
Kordzadeh, N., & Ghasemaghaei, M. (2022). Algorithmic bias: Review, synthesis, and future research directions. European Journal of Information Systems, 31(3), 388–409.
Leonardi, P. M. (2021). COVID-19 and the new technologies of organizing: Digital exhaust, digital footprints, and artificial intelligence in the wake of remote work. Journal of Management Studies, 58(1), 249–252.
Li, S., Liu, Y., & Feng, J. (2023). Who should own the data? The impact of data ownership shift from the service provider to consumers. Journal of Management Information Systems, 40(2), 366–400.
McCartney, S., & Fu, N. (2022). Promise versus reality: A systematic review of the ongoing debates in people analytics. Journal of Organizational Effectiveness: People and Performance, 9(2), 281–311.
Milanez, A., Lemmens, A., & Ruggiu, C. (2025). Algorithmic management in the workplace: New evidence from an OECD employer survey (OECD Artificial Intelligence Papers No. 31). OECD Publishing.
Ngwenyama, O., Rowe, F., Klein, S., & Henriksen, H. Z. (2024). The open prison of the big data revolution: False consciousness, faustian bargains, and digital entrapment. Information Systems Research, 35(4), 2030–2058.
Nyman, S., Benfeldt, O., Gierlich-Joas, M., Bagger, C., Zambach, S., Blicher, A., Mazumdar, S., Nørbjerg, J., Boedker, M., Parra-Moyano, J., & others. (2024). From algorithmic management to data-driven labour organising: A trade union approach to workplace datafication. Scandinavian Journal of Information Systems, 36(1), 1–28.
Parent-Rocheleau, X., & Parker, S. K. (2021). Algorithms as work designers: How algorithmic management influences the design of jobs. Human Resource Management Review, 32(3), 100838.
Rahman, M. S. (2021). The Uber Game: Labor platform capitalism and the algorithmic control of independent workers. Springer.
Rousseau, D. M. (1989). Psychological and implied contracts in organizations. Employee Responsibilities and Rights Journal, 2(2), 121–139.
Ruschemeier, H., & Hondrich, L. J. (2024). Automation bias in public administration: An interdisciplinary perspective from law and psychology. Government Information Quarterly, 41(3), 101953.
Tong, S., Jia, N., Luo, X., & Fang, Z. (2021). The Janus face of artificial intelligence feedback: Deployment versus disclosure effects on employee performance. Strategic Management Journal, 42(9), 1600–1631.
Tursunbayeva, A., Di Lauro, S., & Pagliari, C. (2018). People analytics: A scoping review of conceptual boundaries and value propositions. International Journal of Information Management, 43, 224–247.
Tursunbayeva, A., Pagliari, C., Di Lauro, S., & Antonelli, G. (2022). The ethics of people analytics: Risks, opportunities and recommendations. Personnel Review, 51(3), 900–921.
Zuboff, S. (2015). Big other: Surveillance capitalism and the prospects of an information civilization. Journal of Information Technology, 30(1), 75–89.
Jonathan H. Westover, PhD is Chief Research Officer (Nexus Institute for Work and AI); Associate Dean and Director of HR Academic Programs (WGU); Professor, Organizational Leadership (UVU); OD/HR/Leadership Consultant (Human Capital Innovations). Read Jonathan Westover's executive profile here.
Suggested Citation: Westover, J. H. (2026). People Analytics and Trust: When Transparency Reveals Too Much. Human Capital Leadership Review, 34(4). doi.org/10.70175/hclreview.2020.34.4.4
