Make Dignity Core to Employee Data Use

The amount of employee data that organizations collect is exploding with the digitization of work and with novel digital connections with employees such as workforce wearables (e.g., smart watch, wearable panic button) and mobile devices. We suspect few leaders appreciate the diversity and volume of data about employees that their organizations are amassing. Just during the recent pandemic, employee data has expanded to include vaccination cards and frequent health checks, virtual meeting behaviors, and work-life surveys.

At most large global firms, people and workforce analytics are mainstream initiatives led by HR leaders. But increasingly employee data is being used beyond HR, and it is being combined in new ways. In recent MIT CISR research, we studied dozens of AI projects, many of which included employee data.[foot]In 2019–2020, MIT CISR researchers conducted 100 interviews with participants of 52 distinct AI projects at 48 firms. In 2021, the authors reviewed 22 of the projects that demonstrated significant use of employee data and analyzed them using the lens of the CARE framework.[/foot] For example, one initiative analyzed deidentified employee badge use, affiliated business unit, Wi-Fi device connection activity, and building characteristics to generate insights about building occupancy, and saved the organization millions of dollars in reduced heating and cooling costs.

Innovative uses of employee data, however, can surface uncomfortable insights. Particularly in digital transformation contexts, employee behaviors and knowledge are key to understanding how an organization has historically operated. This understanding can expose opportunities for improvement, and lead to unanticipated outcomes when the organization decides to radically alter or eradicate work tasks moving forward. Such uses of the data can cause tensions and be fraught with gnarly ethical concerns.

Organizations may be tempted to govern employee data by relying on guidance from regulations such as GDPR and HIPAA, which are rooted in personal data privacy and protection. But these laws fall short when it comes to ethical oversight of employee data use. For one thing, employees present with ethical complexity due to the employer-employee relationship and because of their embeddedness in firm operations. Also, MIT CISR research has shown that a regulatory-based perspective is not broad or deep enough to comprehensively oversee the internal and external use of people data; firms need a capability known as Acceptable Data Use (ADU) that includes legal, regulatory, and ethical oversight practices.[foot]B. H. Wixom and M. L. Markus, “To Develop Acceptable Data Use, Build Company Norms,” MIT Sloan CISR Research Briefing, Vol. XVII, No. 4, April 2017, https://cisr.mit.edu/publication/2017_0401_AcceptableDataUse_WixomMarkus.[/foot] ADU also covers values-based oversight, which incorporates the expectations and desires of the organization and key stakeholders such as customers, donors, and suppliers.

An academic article published this year by the first two authors of this briefing suggests that leaders can effectively manage ADU-associated ethical and values-based requirements of employee data use by focusing on dignity.[foot]Dorothy E. Leidner and Olgerta Tona, “The CARE Theory of Dignity and Personal Data Digitization,” MIS Quarterly, Vol. 45, No. 1b, January 15, 2021, https://misq.org/the-care-theory-of-dignity-and-personal-data-digitalization.html.[/foot] In fact, we believe that putting dignity at the center of acceptable data use not only improves ADU management but also allows transformation leaders to reinforce the organization’s regard for its employees.

Understand the Dimensions of Employee Data

We refer to the comprehensive set of interrelated contemporary employee data as 5-W data—data regarding who, what, where, when, and why:

• Who-data describes the employees who are performing work. This has evolved from employee demographics, contact information, health history, and salary and benefits to include social network connections and biometrics from wearables.
• What-data describes employee work activities. This can include online behaviors like internet searches and keyboard actions as well as digitized offline behaviors such as video surveillance, call center audio transcripts, and work logs.
• Where-data discloses employee whereabouts. Where-data such as physical locale and spatial movement has become increasingly more precise with the introduction of workforce wearables, mobile tracking within a building, and smart factories.
• When-data marks the timing of employee activities and related events or outcomes. This can track simple milestones of a workday or work task, or it can reflect a complex time series of events assembled using a variety of sources such as usage logs, mobile devices, sensors, and transactions.
• Why-data historically has represented tacit employee knowledge, including expertise and logic. With digitization, particularly associated with work automation, organizations manage growing repositories about what employees know, how they form judgements and make decisions, and how they feel. For example, AI training data increasingly includes employee feedback about model outcomes so that employee rationale and validation can improve model performance over time.

What is striking about the diverse array of employee data is its combinatorial potential. Not only can organizations combine the 5-W data in myriad ways, but they also can combine employee data with data about customers, operations, products, and other internal subject areas as well as with data sourced externally.

Understand the Use of Employee Data

With 5-W data in hand, the organization can use it to know employees—or to show employees (see figure 1). During digital transformation, both activities are important to pursuing transformation goals; knowing and showing help organizations assess, improve, change, and engineer work.

Knowing employees refers to when an organization uses employee data to understand employee work activities, performance, and behavior and how they relate to desired states or outcomes. We have observed this in AI projects that automated or reengineered manual processes; for example, by knowing what employees are doing and why they performed tasks, a project team can create explicit business rules or train supervised AI models that fuel more efficient processes.

Showing employees happens when an organization uses employee data to communicate insights to employees or to others inside or outside the organization. For example, we have studied many AI projects that compared human and model output to train AI models to be accurate. As anomalies and insights emerge from training, some teams share employee what and why data with the business domain to correct misperceptions or remediate poor practices. Some teams show employees data for motivational reasons; for example, teams might report employee adoption or use of new work tasks so that managers can monitor and manage transformation progress.

Notably, organizations at times share with customers or external partners insights from knowing and showing employees. For example, they might share what they know about employee behaviors to partners who offer services to their employees, such as parking or corporate perks. They might show anonymized employee work practices or outcomes to a benchmark provider that assesses industry-level performance. And they might show employee whereabouts to customers to make service processes transparent.

Evaluate Employee Data Use—Using Dignity

Organizations benefit big as they know and show employees to inform how to improve work practices and delight customers—to the tune of realizing huge cost efficiencies and top-line growth and reaching previously unattainable goals. As organizations enjoy big wins using employee data, however, leaders must ensure that employees are being treated like people, not like objects being managed. In fact, for organizations that establish a North Star of valuing employees, treating employee data use acceptably—by treating employees with dignity—is an excellent way to walk the talk.

Dignity broadly refers to the recognition that human beings possess intrinsic value, are endowed with certain rights, and deserve respect.[foot]Leidner and Tona, “The CARE Theory of Dignity and Personal Data Digitization.”[/foot] Dignity takes three forms: behavioral, meritocratic, and inherent:

• Behavioral dignity: individuals have resources to achieve a life of well-being
• Meritocratic dignity: individuals receive appropriate acknowledgement of their contributions
• Inherent dignity: individuals are treated as worthy of respect regardless of status The behavioral, meritocratic, and inherent forms of dignity each offer a distinct way to evaluate whether an organization supports employee dignity—or threatens it (see figure 2).

Using dignity as a lens, organizations can recognize employee dignity by (1) providing employees with the resources they need to achieve their work, (2) acknowledging excellent performance and achievements, and (3) treating employees with belonging so that they feel accepted and included in organizational matters.

Make Dignity Core to Your Organization’s Employee Data Use

As organizations continue to collect and use 5-W data in new ways, leaders should anticipate ongoing attention to employee dignity as a fruitful part of enterprise data governance processes. Such attention can help organizations treat employees as people at the same time that employees’ work behaviors contribute to the movement regarding automation, task redesign, and future-ready pursuits. To assess your organization’s state of employee data use, ask yourself:

• What 5-W data is in use? How accurate and complete is our 360-degree employee view?
• Are employees aware of 5-W data that is captured on them and do they consent to its use? Do they have options regarding which data is gathered?
• Do employees see their own data, and do they benefit from insights derived from their data?
• Do employees understand how and why their data benefits the organization—and how and why their data benefits them both as part of the organization and personally?