SUBSCRIBE E-NEWSLETTERS AWARDS COLUMNS MULTIMEDIA CONFERENCES ABOUT US RESEARCH

Holding On to the Stars

Are star performers leaving you? Workforce analytics can help you determine why, and what to do about it.

Wednesday, August 14, 2013
Write To The Editor Reprints

Carla, a senior vice president of finance, walked into Rick's office and found him looking really worried. Rick had been doing some amazing things for the organization as its chief marketing officer. He was on the verge of the biggest marketing launch in the company's history, but he seemed to be very worried this day.

"What's on your mind, Rick? Are you having issues with the suppliers again?" asked Carla.

"No, the suppliers are all very aligned," Rick replied. "That said, I just lost another superstar on my team -- Paul is the third so far. This quarter has been really bad for losing good people. I am not sure what it is: Stress? Job satisfaction? Long hours?

"And Paul, like others at the exit interview, apparently said it was just a combination of things," he went on. "There was nothing that I could act on, especially after the fact. I wish I had found some of these red flags sooner than later. I could have taken some preventive action. Now, we need to find someone new, train him or her, get this person up to speed, etc. ... this can take a long, long time."

Have you found yourself in Rick's shoes? In a recent global survey by New York-based Deloitte, titled Talent Edge 2020: Redrafting Talent Strategies for the Uneven Recovery, 71 percent of 376 senior executives surveyed expressed high (43 percent or very high (28 percent) concern about losing critical and high-performing talent. In today's aggressive business environment and tight talent markets, organizations should view their star performers as among their most valuable off-balance-sheet assets. It's no wonder that retaining these assets preoccupies the minds of many executives and HR leaders.

Attrition costs can represent millions of dollars to a medium-sized company each year. Tangible costs include loss in productivity; costs to recruit, hire and train new employees; employee-exit fees; opportunity costs; customer confidence, etc. Intangible costs may also add to the equation, such as employee morale, company brand and reputational issues. According to the Society for Human Resource Management's recent Cost of Turnover report, attrition costs can range from 50 percent to 200 percent of an employee's annual salary. Some studies peg them even higher when they involve the attrition of star-performing employees. Consequently, the stakes are high. Even a nominal reduction in attrition of the best employees can significantly improve a company's top and bottom lines.

Putting the Data to Work

The benefits of getting ahead of attrition issues can be numerous. However, traditional ways of doing so seem to have failed to provide significant results. We assert that there can be three main challenges with traditional approaches, such as those employed by Rick's organization, to address employee attrition.

First, these approaches are mainly reactive. Typically, they consist of analyzing historical HR data and monitoring attrition levels after star performers have left. Second, mitigating actions tend to be expensive. Typical actions involve instituting blanket commitment and motivation-improvement initiatives, and using other methods such as retention bonuses. Such organization-wide campaigns can help increase overall morale and organizational commitment, but can be quite costly. Third, the return-on-investment of mitigating actions is typically not high. The reasons for star performers leaving are different for each individual and are usually complex. Talent managers can have a tough time tailoring organization-wide initiatives to the perceived or stated needs of individual employees. Consequently, managers can find it hard to get enough mileage out of their engagement dollars.

Over the last decade, advanced workforce analytics have begun to find a valuable place in the talent-management domain. With the adoption of data-rich HR enterprise-resource-planning systems, numerous firms now have more relevant, timely and accessible employee data than ever before. Some of the data elements center around employee satisfaction and production levels (e.g., number of high-burn projects, survey scores, corporate perception and commitment to strategy), external stress factors (e.g., commute distance to work, amount of vacation used, travel requirements and work/life balance assessments), individual employee characteristics (e.g., career and life stages, educational background and goal/metric attainment), comparison to peers (e.g., network and interaction with star performers, peer and market compensation levels, supervisor/mentor quality), employee performance (e.g., promotion history and relative compensation level) and so on.

Additionally, with the advent of social networking and "big data," employee blogs, postings, tweets and the like may now be public information and accessible for mining of satisfaction levels. From a macro perspective, organizations also have access to data on economic outlook, national and regional employment rates, salary surveys, geographic employment levels/trends, etc.

As a result, HR leadership at many companies are warming up to the idea of using data analytics to identify actionable insights among complex patterns -- patterns related to individualized employee attrition and other factors that allow for proactive actions to avoid the loss of the most valued team members.

In early 2012, as part of an extensive research effort involving a large time-series-based HR data set -- see Deloitte Consulting's Retention Analytics from the Inside Out -- Deloitte developed an advanced workforce-analytics-driven approach to get ahead of its own attrition. In so doing, the company has taken employee-level data coming from HR (e.g., age and gender), travel expenses (e.g., travel frequency and location), time sheets (e.g., hours worked and vacation time), project financials (e.g., project team composition and project length), as well as market-related information (e.g., unemployment rate and GDP change) and has run various advanced statistical algorithms to try and predict employee attrition in the next fiscal year. The results have been fascinating and, at times, counter-intuitive. A few of the insights that emerged were as follows.

First, while a high amount of travel is generally considered an attrition driver due to the stresses and strains on work/life balance, it was found that a lack, or disproportionately lower amount, of travel for employees with lower tenure, as well as those with higher tenure, seemed to drive attrition. These employee sub-groups enjoyed travel and knew that those who traveled more tended to be more productive, advanced faster and were compensated at higher levels. It was found that the lower-tenured employees were typically looking to show their mettle in their roles and the higher-tenure leaders and executives tended to thrive off of building and growing relationships. Employees who did not travel as much seemed to perceive their careers at the organization as not advancing as they hoped.

Second, while high-burn, high-stress and tight-timeframe projects seemed to wear people down more, employees with disproportionate high and low weekly burn rates were at high risk of attrition (see chart). In addition, once paid-time-off was factored in, a slightly different picture emerged: Employees who took PTO regularly and to the fullest tended to attrite much less compared to their peers who either did not take time off or whose job responsibilities made time off challenging.

Third, employees who repeatedly worked in teams with other star performers tended to leave at a higher rate. One hypothesis for this finding was that the performance and expectation bar was raised so high that peer pressure negatively impacted the actual or perceived performance of an individual.

Alternately, another explanation may be that, when working in teams with numerous star performers, a pecking order can emerge that artificially discourages those lower in that hierarchy, even though all involved are excellent and valued workers.

Finally, employees who regularly blog and/or are active in their respective social networks, both publicly and on internal corporate networks, tend to attrite at a lower rate. Perhaps these individuals have found an avenue to express themselves professionally and establish their professional identities?

Cutting Attrition

While these findings are interesting and cause much discussion among talent-management professionals, what makes workforce-analytics techniques so promising is that they can allow decision-makers to understand the drivers that are specific to an individual versus more traditional macro trends. This can allow management to be more precise in identifying red flags and employees disproportionately at risk to leave and with sufficient lead time to potentially improve outcomes. We have seen these techniques allow companies to be more proactive rather than reactive, and to get ahead of the problem.

Newsletter Sign-Up:

Benefits
HR Technology
Talent Management
HR Leadership
Inside HR Tech
HRENow
Special Offers

Email Address



Privacy Policy

When it comes to turnover and attrition, leadership actions should matter. With a focus on utilizing, engaging and developing employee skills, satisfied employees may likely stay faithful to an organization. In addition, continuously recognizing and rewarding success and hard work translates into a more committed workforce. Lastly, returns on communication are underrated: Companies that communicate effectively and transparently are more likely to engender trust, strengthen employee job satisfaction and retain top workers.

Because workforce-analytic-driven interventions are targeted to specific employees, the ROI tends to be much higher. It costs significantly less to address the specific needs of one individual or subgroup, or sub-region, as opposed to addressing the general needs in an organization through "spray and pray" retention campaigns.

Such employee-specific targeted mitigation and intervention strategies require creativity, flexibility and commitment. For example, a mid-tenure employee who is on high-burn projects and on projects involving a lot of travel can be put on internal projects periodically with positive messaging about why it's being done and how valued the individual is to the company. Other intervention strategies could be to allow lower-tenured employees who are not traveling as much to be placed on assignments with travel or develop targeted actions to encourage employees on high-burn projects to take more time off between projects so they can sustain the pace -- with positive and encouraging messaging by way of explanation.

As an unintended consequence, while the model at Deloitte was aimed at predicting who was most likely to leave, it also revealed a couple of very interesting corollary findings. First, not all voluntary attrition was bad. Some of it was in the best interests of the organization and the employee specifically when the predicted departure was a result of poor performance. In these cases, the preferred action was no action. Second, while the model was built to predict who was most likely to leave, the model also and inversely revealed some characteristics that are important for driving employee retention.

During the research, as drivers of retention came to be known in terms of direction and magnitude, Deloitte looked for ways to incorporate these findings into its retention programs and strategies. For example, managers were encouraged to watch for employees who were traveling extensively without taking much vacation time. When such red flags emerged, managers would take action to redefine travel and project schedules. Similarly, when employees reached a certain time threshold at their current positions, rotation and training programs would be triggered and redeployment strategies initiated. While some drivers were not directly actionable (for example, high unemployment rates were associated with lower voluntary attrition) many more were directly or indirectly actionable.

While workforce analytics shows great promise, it is not a panacea. Organizations should be equipped to deal with false positives emerging from the data. With organizational discretion, managers need to review the lists of those most likely to leave and interpret the reason codes that emerge. False negatives can also arise from model error or statistical anomaly. No model is 100 percent predictive; all findings should be scrutinized and reasonability checked. While targeted intervention campaigns do yield a higher ROI, caution should be used so intervention campaigns are perceived as fair and balanced by all employees.

In summary, attrition management for star-performing employees can be a big challenge, a big HR cost area and a big drain on the productivity of an organization. As we heard in the introduction, almost everyone has been in Rick's shoes at one point or another. Evolving leading practices in talent management call for organizations to be proactive and to identify red flags that are specific to each at-risk-to-attrite individual. Most importantly, a sound and well-thought-out implementation strategy should be developed. It has been shown that by doing these things, star employees can be satisfied and retained at higher rates, which results in significant financial benefit to a company. Cutting-edge talent-management techniques using workforce analytics can take you on a competency development journey that should be taken seriously -- the benefits can far outweigh the costs.

To all you Ricks out there, good luck in your respective product launches and in retaining your star talent!

Amel Arhab is a manager with Deloitte, specializing in the development and application of predictive analytics and business intelligence for the financial services and insurance industries; John Houston is a principal with Deloitte Consulting and its national practice leader for advanced analytics and predictive modeling; Vishwa Kolla is a senior consultant in Deloitte's advanced analytics and modeling practice; and John Lucker is Deloitte's global advanced analytics and modeling market leader and a leader within the U.S. Deloitte analytics practice.

Copyright 2014© LRP Publications