We’re Already Living in the Future of Talent Analytics

We’re Already Living in the Future of Talent Analytics

Recently at the Harvard Business Review, management professor Thomas H. Davenport asserted that HR “is right up there with the most analytical functions in business—and even a bit ahead of a quantitatively-oriented function like finance.” Davenport backs this claim with findings from a global survey of senior managers, directors, and VPs at large companies by Oracle, on which he collaborated. The survey found that many HR leaders are well-versed in using data and predictive analytics to make talent management decisions:

  • 51% of HR respondents said that they could perform predictive or prescriptive analytics, whereas only 37% of Finance respondents could undertake these more advanced forms of analytics.
  • 89% agreed or agreed strongly that “My HR function is highly skilled at using data to determine future workforce plans currently (e.g. talent needed),” and only 1% disagreed.
  • 94% agreed that “We are able to predict the likelihood of turnover in critical roles with a high degree of confidence currently.”
  • 94% also agreed that, “We have accurate, real-time insight into our employees’ career development goals currently.”
  • When asked “Which of the following analytics are you using?” “artificial intelligence” received the highest response, with 31%. When asked for further detail on how respondents were using AI, the most common responses were “identifying at-risk talent through attrition modeling,” “predicting high-performing recruits,” and “sourcing best-fit candidates with resume analysis.”
  • These findings suggest that the analytics transformation in HR is farther along than you might have thought, with the caveat that the survey respondents were from companies with $100 million in revenue or more, and are thus more likely to have the capacity to deploy new techniques and technologies that may be out of reach for smaller organizations. It should come as no surprise that more and more companies are adopting AI and analytics into their HR functions; what’s new in this survey data is that HR functions are becoming increasingly confident in the maturity and capability of their analytics programs.

    In terms of where companies are deploying talent analytics, Oracle’s findings track with what we have seen elsewhere: The lowest-hanging fruit is in predicting turnover, while there’s also a lot of promise in AI-powered recruiting, predicting performance, and career pathing. The focus on attrition makes sense, as employees who quit often time that decision to leave around predictable life and career events and drop lots of hints about their plans beforehand.

    If you can use data to detect these warning signs and head off unwanted departures, that can save your organization considerable amounts of money. IBM CEO Ginni Rometty made headlines earlier this month when she told attendees at CNBC’s @Work Talent + HR Summit that IBM’s AI technology was able to predict which workers were planning to quit with 95 percent accuracy:

    IBM HR has a patent for its “predictive attrition program” which was developed with Watson to predict employee flight risk and prescribe actions for managers to engage employees. Rometty would not explain “the secret sauce” that allowed the AI to work so effectively in identifying workers about to jump (officially, IBM said the predictions are now in the 95 percent accuracy “range”). Rometty would only say that its success comes through analyzing many data points.

    “It took time to convince company management it was accurate,” Rometty said, but the AI has so far saved IBM nearly $300 million in retention costs, she claimed.

    But predicting turnover with enough accuracy to add value may not require IBM-level AI capabilities. A new study from Peakon finds that employees begin showing clear signs of wanting to quit a full nine months before they pull the trigger on their resignation. A big-data study drawn from over 32 million employee survey responses in 125 countries, the Peakon report points to several key indicators of attrition that show up months in advance: declining engagement and loyalty, as well as dissatisfaction based on unchallenging work, an inability to discuss pay, an unsupportive manager, and the lack of a clear path to advancement in the organization.

    In a recent interview with David McCann at CFO, data scientist Jon Christiansen notes that it’s much easier to predict who will stay than who will leave, but highlights a few indicators that consistently point toward a greater likelihood that an employee will quit, such as whether the employee feels that their performance is evaluated fairly or that they have control over their workday. Other signs include an employee avoiding conflict, siloing themselves, focusing excessively on rewards over the common goal of the organization, and facing either too much or too little pressure at work.

    The advantage for a company like IBM, which continues to invest heavily in AI, is that it can delegate the detection of these patterns to an algorithm. Predicting quits was the first area the tech giant’s HR function focused on when deploying AI, IBM’s chief human resources officer Diane Gherson explained to Jena McGregor at the Washington Post:

    IBM had already been using algorithms and testing hypotheses about who would leave and why. Simple factors, such as the length of an employee’s commute, were helpful but only so telling. “You can’t possibly come up with every case,” Gherson said. “The value you get from AI is it doesn’t rely on hypotheses being developed in advance; it actually finds the patterns.”

    For instance, the system spotted one software engineer who hadn’t been promoted at the same rate as three female peers who all came from the same top university computer science program. The women had all been at IBM for four years but worked in different parts of the sprawling company. While her manager didn’t know she was comparing herself to these women, the engineer was all too aware her former classmates had been promoted and she hadn’t, Gherson said. After the risk was flagged, she was given more mentoring and stretch assignments, and she remains at IBM.

    IBM is also using its Watson AI for other talent-related purposes, such as learning and development or career pathing, Carrie Altieri, IBM’s vice president of communications for people and culture, noted in a recent interview with Riia O’Donnell at HR Dive:

    AI has been a driving force of innovation for IBM’s HR team. Cognitive talent alerts mine for patterns; it searches for employees who’ve been in a job longer than usual (which could signal flight risk) and can determine whether they need more training to move up. …

    AI also can personalize learning and development for each job role and lead the way in making learning a central aspect of a company’s culture. Altieri said that more than 45,000 learners are visiting IBM’s learning platform every day and 98% of employees access it each quarter. While the company requires 40 hours of learning per year, staff average around 50 hours, regardless of tenure. Learning is a huge part of the culture at IBM, she explained, and the new system gives managers the tools to have more intentional discussions with staff.

    And like other tech companies experimenting with these technologies, IBM is not only deploying its AI capabilities internally, but also selling them as a service to other organizations. Last November, the company announced the launch of IBM Talent & Transformation, a new business venture offering AI skills training in addition to services that “harness the power of AI personalization to guide employees in developing skills and pursuing opportunities to grow within the company.”

    Legion Aims to Solve the Scheduling Problem with Data

    Legion Aims to Solve the Scheduling Problem with Data

    In sectors like food service and retail, where front-line employees work hourly and customer traffic is highly variable, the conflict between businesses’ need for flexible staffing and employees’ desire for predictable hours and incomes has led to increased labor activism and efforts to regulate variable scheduling. Even though most employees with variable schedules don’t have a problem with them, they can be a hardship for low-income employees struggling to make ends meet, or parents trying to schedule around the needs of their children. The controversy has led to some major retailers dropping the practice of “on-call” scheduling.

    Fortunately, a growing number of technological solutions are coming to market to help organizations set and communicate schedules in ways that are more predictable and less disruptive to their employees. The latest of these is a startup called Legion, which recently raised $10.5 million in funding for its platform. Founder Sanish Mondkar tells TechCrunch’s Matthew Lynley that he hopes to use big data to crack the challenge of intelligent scheduling once and for all:

    The startup uses large amounts of data, all the way down to the weather near a store, to try to predict how busy it will be and how to intelligently staff that store and prepare for the foot traffic. It also works to sort out the best possible schedule for each employee, whether they want to work a regular shift at the same hours or vary from week to week and trade shifts a lot. The company is rolling out with Philz, one of Silicon Valley’s favorite coffee projects, to try to prove out such a concept. …

    Read more

    The Future of Law Is Digital?

    The Future of Law Is Digital?

    While most of the public discourse about automation focuses on robots displacing manufacturing employees, other emerging technologies like artificial intelligence and machine learning also stand to automate the work that many knowledge workers do as well—in fact, they already are. One area where AI is poised to have a big impact is the legal sector, where it might significantly disrupt the nature and cost of lawyers’ work. In the Atlantic’s April issue, Jason Koebler takes a closer look at the emergence of robo-lawyers, focusing on several ways AI is already being applied to the practice of law in new and potentially unsettling ways, such as using statistical data to predict the likely outcomes of court cases:

    Beyond helping prepare cases, AI could also predict how they’ll hold up in court. Lex Machina, a company owned by LexisNexis, offers what it calls “moneyball lawyering.” It applies natural-language processing to millions of court decisions to find trends that can be used to a law firm’s advantage. For instance, the software can determine which judges tend to favor plaintiffs, summarize the legal strategies of opposing lawyers based on their case histories, and determine the arguments most likely to convince specific judges. A Miami-based company called Premonition goes one step further and promises to predict the winner of a case before it even goes to court, based on statistical analyses of verdicts in similar cases. “Which attorneys win before which judges? Premonition knows,” the company says.

    Read more

    A Lesson in Using Predictive Analytics to Drive Retention

    A Lesson in Using Predictive Analytics to Drive Retention

    When we talk to HR leaders about predictive analytics, the first thing they usually want to do with this new advanced tool is improve retention. That’s definitely easier said than done, especially if you want the project to actually drive results, instead of just being an interesting research topic. Aliah Wright at SHRM highlights the success story of one organization that had a strong need to retain its highly skilled employees and used predictive analytics to help meet that goal:

    When a top employee at the Anderson Center for Autism, a private school in Staatsburg, N.Y., handed in her resignation, the school’s HR department was expecting her. The HR staff had been using a predictive analytics program to help them gauge retention. “The software is so good that we were developing a retention plan for her as she was preparing to resign,” said Gregg Paulk, director of information technologies for the 92-year-old nonprofit organization. After HR staff spoke with her, “she actually rescinded her resignation,” he added. …

    In 2001, the school undertook a new technology initiative spurred and funded by the No Child Left Behind legislation. Using Ultimate Software’s UltiPro, Paulk said the company “grew … and kept head count flat, reduced paper [processes] by 95 percent, and increased the time spent on employee development by 30 percent. The software also allows staff to manage time and attendance from anywhere [and yields] improved reporting and compliance.

    “The software also helped us avoid the loss of key talent with predictive tools. It’s really powerful, and it’s astonishing the results we’ve seen,” Paulk said. “[The tools] helped us understand our challenges and put the puzzle pieces together.”

    It looks like Anderson has done a couple of things really well, which makes it a great example of how to apply analytics most effectively.

    Read more

    Salesforce Gets in the AI Game as Microsoft Chases Slack

    Salesforce Gets in the AI Game as Microsoft Chases Slack

    Artificial intelligence is beginning to make its presence felt in every corner of the working world. AI-powered chatbots are the future of workplace communication platforms like Slack, while the HR technology company FirstJob’s digital recruiting assistant Mya uses AI to streamline the hiring process and improve candidate experience. Now Salesforce is also introducing AI enhancements to its sales software, the Wall Street Journal reports:

    Called Einstein, the new offering is a set of online AI services designed to automate tasks, predict behavior and spotlight relevant information. … If Salesforce can integrate AI into its applications, the San Francisco, Calif., cloud-software provider may get a jump on competitors, said Tom Austin, vice president at research firm Gartner Inc. “There are no simple, easy applications today to buy that really work,” he said. “This is hard stuff still.”

    Beyond offering AI-equipped applications, Salesforce is joining companies such as IBM, Microsoft and Google in making AI services available for other companies to build into their own applications. Prior to their efforts, AI was limited to companies that could afford to hire data scientists and build large computing facilities. Providing access to AI software as a cloud-computing service over the internet lets companies tap the technology with a smaller investment.

    Salesforce said Einstein, which eventually will be built into a range of the company’s products, can better predict which customers are likely to buy products, and recommend which consumers a sales person should contact. It analyzes data stored on Salesforce’s servers such as customer information, email, calendar entries, and social-media posts to learn how specific consumers are likely to behave.

    Another heavyweight making waves in this space is Microsoft. hot on the heels of its acquisition of LinkedIn, the legacy tech giant is building a direct competitor to Slack, which it also considered trying to buy earlier this year, Mehedi Hassan recently revealed at MSPoweruser:

    Read more

    Walmart Tries Out a New Scheduling System

    Walmart Tries Out a New Scheduling System

    In response to persistent calls from employees for more predictable hours, and in order to improve customer service, Walmart is piloting a new version of its scheduling software designed to better predict customer traffic, Reuters reports:

    The system, called Customer First Scheduling, was launched in all of Wal-Mart’s 650 small-format Neighborhood Markets in the last week of July with plans to eventually roll it out across the entire U.S. store network, although the company gave no timeframe. …

    The electronic system can prioritize scheduling for peak shopping hours by taking into account foot traffic and sales data from every department in each store. Staff are then allocated to the remaining shifts in order of importance. Wal-Mart began last year to try and improve customer service with faster checkouts and better-stocked shelves. The new system also aims to give employees more certainty over shifts and should cut down on the need to schedule employees on short notice. …

    The new system allows some workers to have a fixed schedule with the same hours and days for up to six months. Those with unfixed schedules will only be slotted to work when they say they are available and will not be expected to be available on short notice. Currently, Wal-Mart managers allocate hours within the times employees say they are available to work.

    The change has been in the works for some time, the Wall Street Journal’s Sarah Nussauer observes:

    Read more

    Hail to the Algorithim: The Wharton People Analytics Conference

    Hail to the Algorithim: The Wharton People Analytics Conference

    Earlier this month, I attended the Wharton People Analytics Conference in Philadelphia, where start-ups, thought leaders, and practitioners networked, shared their work and debated the future. The event was by turns heady (e.g., a panel with Wharton Dean Geoffrey Garrett and Byron Auguste of Opportunity@Work discussing “the why of work”), insightful (e.g., Google’s recent research on teams), and tactical (e.g., solutions profiled by recruiting startup HiQ and other talent analytics vendors).

    The conference had many highlights, but here are just a few that stood out to me:

    Daniel Kahneman and the Limits of Predictability

    In a talk with Dan Pink, psychologist and behavioral economist Daniel Kahneman shared his thoughts about the evolution of talent assessments and behavioral economics. Of interest were his comments about the limits of our ability to predict performance. Though he argued that many failed predictions are due to noise, not bias, he wondered whether our ability to predict performance would continue to improve or if we’ve hit a limit. Maybe there are some things we just can’t model out effectively.

    In some ways, that may be true for individual performance. But within CLC, we’ve found that more organizations do a very poor job of predicting performance in a more collaborative, team-based environment. Indeed, two thirds of enterprise contributors—our term for high performers—are misidentified. (CEB Corporate Leadership Council members can read more about enterprise contributors and how to develop them here.)

    For talent analytics professionals in general, even if all noise cannot be eliminated from performance models, I would argue that these models still need to be updated from time to time to keep up with the pace of change. The discussion also raises interesting questions about the point of talent analytics, which isn’t just about increasing predictability, but rather about increasing it within a certain timeframe or separating signal from noise when dealing with massive infusions of information (large numbers of job applicants, for example).

    Read more