cornell HR review Sieving through the Data to Find the Person: HR’s Imperative for Balancing Big Data with People Centricity John Lipkin With “big data” and “analytics” atop human resources (HR) professionals’ dictionaries, it is no wonder that some are calling it time to think of employees as data points1 and to scientifically make people decisions.2 These beget horrific images of what many employees already believe HR promotes: incessant change and downsizing solely for profit maximization. Yet, for HR to genuinely transition into the world of data-driven people solutions, it must leverage its roots in employee advocacy, understanding, and development.3 To best do this, HR must undertake three actions. First, HR can ease into people analytics, using the necessary time and effort to gain employee buy-in. Second, HR should stress the objectivity of data-driven decision making. Third, HR practitioners must exhibit empathy for those affected by such decisions. The Case for Data-Driven People Solutions With statistical methods and big data usage entrenched in nearly all other facets of business – from predicting supply and output in supply chain, to predicting a range of financial outcomes in finance, to using mass swaths of information for predicting customer preferences in marketing – HR has been slower to embrace its usage in practice. Analytics falls into one of three categories: descriptive (what happened), predictive (what could happen), and prescriptive (what should happen).4 Utilization of these interwoven statistical approaches helps to clarify historical trends, find correlations and patterns among variables, look into the future, mitigate legal risk, and maintain objective decision-making.5 Furthermore, the past 20 years have demonstrated that HR analytics can provide a return on investment (ROI) that influences business leaders to support HR agendas and align with organizations’ strategies. For example, Sears used causal pathway modeling to predict that a 5 percent increase in employee engagement drove a 0.5 percent increase in revenue, which has led to an additional $200 million annually.6 Another example hails from organizational design and effectiveness. Using sensors that track employee movement, Bank of America (BofA) found that call-center groups where employees did activities together (e.g. take breaks) performed better than less cohesive ones. A simple action - require all call-center employees take breaks as groups – increased agents’ efficiency by 10 percent.7 While these examples set the stage for analytics’ worth in © 2015 Cornell HR Review driving business outcomes, the following sections focus on data-driven, people-centric implementation. Easing Into Data-Driven People Decisions with Smaller, Quick Wins Change management guru John Kotter famously advocated any successful change require eight steps. HR must not forget that implementing data-driven solutions is indeed a change for both employees and the organization. In implementing data-driven solutions, one particular step – creating smaller, quick wins – allows HR to gain buy-in through empirical results. Creating smaller, quick wins can take many forms. First, HR can utilize smaller employee populations before rolling out the solution to the entire targeted group. Second, HR might consider which areas of HR it chooses to first use analytical approaches. For example, areas like recruiting and retention cause much less strife than does performance management. Third, HR can ease into the types of data it analyzes. Smaller Populations: Sometimes the grand rollout is not always the best decision. Unveiling the data-driven solution to a smaller population first allows the company to more nimbly adjust its algorithm, its implementation strategy, or scrap the idea altogether. Google, according to SVP of People Operations Laszlo Bock, rarely rolls out a people analytics decision without first using a small pilot group.8 Other possibilities include unveiling the solution to an entire function or an entire location before it goes company-wide. These procedures enable HR to not only change when needed but also compare results of the targeted group to other groups. By seeing the successes of the targeted group, employees will be less likely to push back, when a larger implementation occurs. Functional Areas of HR: Although many organizations aspire to predict future performance of their employees, companies might consider using data-driven solutions to focus on other important HR areas before addressing ones that employees consider more divisive. Areas like recruiting – predicting which attributes, however zany they may seem, lead to success in organizations – and retention – identifying the drivers of turnover and intervening to improve upon those – are much less likely to cause pushback. With wins in less contentious areas, HR departments can move to implement data-driven solutions in areas less contentious with employees. Types of Data: Employers can collect varying degrees of data. These include anything from demographic information, performance information, and surveys to newer areas of data capture like wearables, machine learning, language processing, and email aggregators. Besides potential legal ramifications, employees may feel uncomfortable with the new generation of big data. By creating smaller, quick wins with the more accepted data, HR can prove to employees the value of these newer forms of data, if and when it determines. 2 © 2015 Cornell HR Review Communicating Objectivity and Transparency of Data-Driven Solutions A mere 5 percent of employees understand their employers’ business strategies.9 This alarming fact highlights just how poor companies communicate with their employees. Thus, a proper communication strategy can significantly aid how employees receive data- driven people decisions. Data-driven people decisions have a unique lever that other decisions lack: data is evidence-based. Thus, HR departments can utilize this attribute to mollify employees’ discontent with data-driven decisions. Objectivity: According to Laurie Bassi from McBassi and Company, “HR analytics is an evidence-based approach for making better decisions on the people side of the business.”10 An evidence-based approach relies on facts and refrains from subjectivity.11 HR departments must take advantage of the objectivity that HR analytics and data-driven decision making provide. HR can be especially powerful in presenting such objective information to employees in fields that more frequently work with data, such as the engineers at Google. By stressing objectivity, employees will more likely accept the decision. Communication: Transparency: Fostering a culture of transparency can abet data-driven people solutions. In a study across generations, transparency comes up as one of the traits employees value most in their bosses.12 Thus, to mitigate the challenge to data-driven people decisions, companies should transparently present the data to their employees. For example, Google holds weekly “all-hands” meetings, where everything is shared with employees and employees can ask any question.13 This culture of transparency fosters the trust that in turn leads to buy-in for the data-driven solutions. Communication: Storytelling: In order to create buy-in in the organization, HR must better communicate the story that the data depicts. Doing so has been one of the most neglected aspects of the big data revolution in HR.14 Finding better ways to visualize the data, provide the business case behind it, and sell the importance of interventions will help gain buy-in from employees. Empathy HR still holds a very important role as both businessperson and employee advocate. When managing and implementing data-driven decisions, HR must stay true to this unique responsibility. It is especially important to demonstrate empathy for the employee when decisions threaten her/his comfort zone. Three considerations demonstrate empathy and help balance the data-driven people decision with people-centricity: First, companies can incorporate employees into the analytics process. Second, make sure establish the benefits for the employee. Third, the data should not make the decision. Incorporate Employees: Companies can achieve the dual-result of leveraging employees’ insights and creating buy-in for data-driven decisions by incorporating their employees in the analytics process. Companies can solicit ideas from employees for business problems to analyze. They can further reward the employees who submitted a business case that eventually gets analyzed and intervened with. Regarding analytics and performance 3 © 2015 Cornell HR Review management, companies can invite employees to suggest traits in managers that they value. This might engender a useful variable that the company had never previously considered. Actions like these foster engagement and buy-in. Benefits for Employee: Most data-driven people decisions are designed to help both the business and the employee. They provide insights that lead to targeted interventions to improve a situation. For example, employers can use network analytics to find out who in the company is not communicating with colleagues as much as expected. This is to coach her/him and her/his team to better communicate. Similarly, using analytics to determine which employee groups have a high health risk seeks to implement solutions to improve their health outcomes. Google’s famed Project Oxygen “was always meant to be a developmental tool” and not one tied to performance appraisals.15 Similarly, when Shell managers communicate to employees their current estimated potential (CEP), the next step in the process is to jointly work to improve – or maintain – the CEP.16 These all are intended to help employees. Don’t Rely Solely on Data: Economist Sendhil Mullainathan playfully advised to let data have a seat at the table instead of letting it be the table.17 This holds serious merit. HR departments cannot rely solely on data to make decisions. As Professor Peter Capelli said, “For me, a…concern is that these systems are likely to produce companies…with homogenous workforces.”18 With these and other issues, such as business conditions changing faster than a model can account for, it is imperative that HR gather insights from the data and make informed decisions based on those. Doing so will gain trust from employees. Conclusion In order to balance data-driven people decisions with people-centricity, HR must take actions to ensure that both the business and employees benefit. Three of these include easing into analytics with smaller, quick wins, communicating the objectivity of data, and demonstrating empathy throughout the process. Consider that after the failed implementation of the algorithm to predict engineers’ success, Google’s VP of People Analytics Prasad Setty did not just scrap the whole project. He came back the next year, instead brandishing better communication and more empathy.19 ℵ John Lipkin received his Master of Industrial and Labor Relations from Cornell University in May 2015. He is currently working in HR Analytics and Compensation with Cigna. Originally from Fort Lauderdale, Florida, John worked in government relations and refugee resettlement before attending Cornell. 1 Mims, C. (2015). In ‘People Analytics,’ You’re Not a Human, You’re a Data Point. The Wall Street Journal. Retrieved April 5, 2015, from http://www.wsj.com/articles/in-people-analytics-youre-not-a- human-youre-a-data-point-1424133771. 2 Bersin, J. (2013). Big Data in Human Resources: Talent Analytics (People Analytics) Comes of Age. Forbes.com. Retrieved April 5, 2015 from http://www.forbes.com/sites/joshbersin/2013/02/17/bigdata-in- human-resources-talent-analytics-comes-of-age/. 3 Leonard, B. (1996). Balancing Business Partnership with Employee Advocacy. HR Magazine, 41:1, p. 89. 4 Human Capital Analytics @ Work (2014). The Conference Board. 4 © 2015 Cornell HR Review 5 Sullivan, J. (2014). 11 Good Reasons That HR Really Needs Predictive Analytics. TLNT.com. Retrieved April 9, 2015 from http://www.tlnt.com/2014/10/06/11-good-reasons-that-hr-really-needs-predictive- analytics/. 6 Rucci, A., Kirn, S., and Quinn, R. (1998). The Employee-Customer-Profit Chain at Sears. Harvard Business Review. Retrieved April 14, 2015 from https://hbr.org/1998/01/the-employee-customer-profit- chain-at-sears/ar/1. 7 Overby, S. (2013). HR Departments Invaded By Data Scientists. CIO. Retrieved April 14, 2015 from http://www.cio.com/article/2383195/business-intelligence/hr-departments-invaded-by-data- scientists.html?page=2. 8 Bock, L. Comments at Wharton People Analytics Conference (2015). 9 Gochman, I. and Storfer, P. (2014). Talent for Tomorrow: Four Secrets for HR Agility in an Uncertain World. People and Strategy, 37:2. 10 Bassi, L. (2011). Raging Debates in HR Analytics. People & Strategy, 34:2, p. 16. 11 Pfeffer, J. and Sutton, R.I. (2006). Evidence-Based Management. Harvard Business Review. Retrieved April 14, 2015 from https://hbr.org/2006/01/evidence-based-management. 12 Myths, Exaggerations, and Uncomfortable Truths: The real story behind Millennials in the workplace (2015). IBM Institute for Business Value. 13 Bock, L. (2011). Passion, Not Perks. Thinkwithgoogle.com. Retrieved April 14, 2015 from https://www.thinkwithgoogle.com/articles/passion-not-perks.html. 14 Gibson, C., Ziskin, I., and Boudreau, J. (2014). What Is the Future of HR?. workforce.com. Retrieved April 14, 2015 from http://www.workforce.com/articles/20179-what-is-the-future-of-hr. 15 Garvin, D. (2013). How Google Sold Its Engineers on Management. Harvard Business Review, pp. 74- 82. 16 Ferrarie, K. (2005). Processes to Assess Leadership Potential Keep Shell’s Talent Pipeline Full. Journal of Organizational Excellence, pp. 17-22. 17 Mullainathan, S. Comments at Wharton People Analytics Conference (2015). 18 Pearlstein, S. (2014). People Analytics: ‘Moneyball’ for human resources. The Washington Post. Retrieved April 14, 2015 from http://www.washingtonpost.com/business/people-analytics-moneyball-for- human-resources/2014/08/01/3a8fb6ac-1749-11e4-9e3b-7f2f110c6265_story.html. 19 Personal communication with Prasad Setty on April 11, 2015. 5