Wise Practitioner – Predictive Workforce Analytics Interview Series: Lisa Disselkamp and Tristan Aubert at Deloitte
By: Greta Roberts, Conference Chair, Predictive Analytics World for Workforce 2016
By: Greta Roberts, Conference Chair, Predictive Analytics World for Workforce 2016
In anticipation of their upcoming Predictive Analytics World for Workforce conference co-presentation, Predictive Analytics Unlocks Sustainable Cost Reduction In Hourly Workforce, we interviewed Lisa Disselkamp, Director at Deloitte, and Tristan Aubert, Senior Consultant – Advanced Analytics & Modeling at Deloitte. View the Q-and-A below to see how Lisa and Tristan have incorporated predictive analytics into the workforce of Deloitte. Also, glimpse what’s in store for the new PAW Workforce conference.
Q: How is a specific line of business / business unit using your predictive decisions? How is your product deployed into operations?
A: [Lisa Disselkamp] Our tool is being used to evaluate operational and financial impact of regulatory changes to exempt labor status by finance, HR and operations. (This is really for 2016.)
Our tool is being used by finance and HR to evaluate compensation policy costs based on operational issues of unique units in their business. They are looking at the necessity of pay programs based on what drives labor demand and employee supply. Based on the analysis, even high cost pay programs may be acceptable based on operational requirements, and low cost programs may stop being overlooked because operationally they are unnecessary.
Decision makers are leveraging the tool to evaluate when and what to make changes to job assignments to reduce cost and maintain necessary front line activities at an acceptable level given the specific situation.
In 2016 our tools will be instrumental in responding to changes in the salary threshold for exempt employees resulting from proposed changes from the Department of Labor. Employers of all types will need to predict the potential cost of converting employees to hourly and the capacity to provide adequate labor hours under more constrained conditions to run their business.
A: [Tristan Aubert] The businesses would use the tool to identify units/teams with higher than expected overtime, understand the drivers of overtime, and make use of this data to inform the actions they would like to take to address this issue before it occurs.
The tool itself does not provide recommendations, rather it provides additional clarity into the causes of overtime and predicts – based on past history – which units are most likely to be prone to overtime and why. For this tool to be used most effectively it needs to be paired with strong domain knowledge into how to best mitigate the drivers that cause the issue and with the business knowledge to determine what areas are in need of attention. Some units may naturally be more prone to overtime – (provide example) – though this does not necessarily mean the functioning should be changed.
Q: If HR were 100% ready and the data were available, what would your boldest data science creations do?
A: [Lisa Disselkamp] HR would be able to forecast how pay policies and schedules are driving labor cost, productivity and revenue. People drive the business and the bottom line and having a more clear picture of how compensation motivates workers and how schedules drive attendance, recruitment and retention would enable them to position workers based on schedule fit, worker skills and cost. They would also help managers influence the daily situations where pay policy and work opportunity converge and potentially inflate labor spending or upset employees.
A: [Tristan Aubert] The right data would mean finding the best approach to manage a person based on their skillset and place in their career to match them with the best opportunities within a firm. I think it’s fairly safe to say that if people enjoy what they do, they will be more productive, motivated and generate better long-term results for their employers. The right data would help define what employees do enjoy about their work situation – how many hours they work, when they work, the predictability and stability of the hours they work, work that includes the activities they want to perform and skills they want to build, all go into job satisfaction. So a comprehensive approach to aligning people to what they most enjoy and are suited for would be the boldest data-science creation.
Q: When do you think businesses will be ready for “black box” workforce predictive methods, such as Random Forests or Neural Networks?
A: [Tristan Aubert] When businesses have understood the implications of using black-box models and determine where it is appropriate to use and where it is not. There are no major technical hurdles to creating black-box models however, human resource challenges are not easily ‘optimized’ in the manner that engineering challenges are. Furthermore, it must be appreciated by end users that models are not infallible or necessarily fair –they tend to reflect pre-existing biases – and there is risk involved in this, doubly-so with black-box models where the mechanics are not well understood.
A: [Lisa Disselkamp] It’s going to take time and practice. Organizations are going to have to lay out a plan that includes incrementally changing the way organizations operate. It could take many rounds of change to build up to these methods. Readiness will come when they have developed the confidence and ability to incorporate predictive analytics into their decision making, not use it to replace human decision making. They also need to be skilled in modeling and testing to validate these intelligent systems are leading them to the proper conclusions. Readiness is not just about conversion but about using tools such as these to strengthen business processes and decision making. The decision making processes and the data must be sound before these tools are deployed. Readiness may involve developing or hiring for the right skill sets which include knowing how the business will be impacted by these predictive tools and managing the transformation to a data driven model.
Q: Do you have suggestions for data scientists trying to explain the complexity of their work, to those solving workforce challenges?
A: [Lisa Disselkamp] Take the time to show them how using the wrong data or not asking the right questions can give a false answer; map out how visualizations are produced so that they are not fooled by charts and graphs and understand what must go on behind the scenes with the mathematical modeling.
A: [Tristan Aubert] If you can’t explain it simply, you don’t understand it well enough – Einstein’s maxim is extremely applicable in this case. A good starting point is to always try and establish a clear link between what you are doing as and the problem at hand. Usually, it is advisable not to dive into the technicalities of the work you are doing but rather to explain – in jargon-free terms – why a particular activity is necessary to arrive to the solution.
Q: What is one specific way in which predictive analytics actively is driving decisions?
A: [Lisa Disselkamp] Making decisions based on data isn’t new. What is new is that predictive analytics gives leaders more confidence to make not only bold moves, but measured moves that are based on solid data and give leaders greater confidence in sustainability once a decision is made. Making a change to an outdated pay policy is easy, getting the outcome you desire and having that “stick” are where predictive analytics can refine and bolster decision making.
A: [Tristan Aubert] Workforce retention models are actively used by analytics-savvy firms to improve on their ability to retain their at-risk talent. As a result of being able identify which segments of the workforce are most at risk of departing, they are able to make informed decisions on how to pursue those individuals deemed critical to the workforce.
Q: How does business culture, including HR, need to evolve to accept the full promise of predictive workforce?
A: [Tristan Aubert] A few things need to happen. First, data generating systems need to be designed with analytics in mind – without good data no insights can be uncovered. Second, business culture needs to understand how to make use of the insights driven by analytics – this means both determining where analytics fits in their day-to-day decision process as well as understanding the nuances and limitations of analytics. Finally, businesses need to develop an imagination of the types of questions that can be addressed by analytics – this will mean going beyond known problems and starting to try to uncover unknown-unknowns.
A: [Lisa Disselkamp] To do that, HR leadership needs to be willing to go beyond superficial changes and have the courage to fundamentally change how their organization operates. Systems, processes and policies may need to be completely abandoned in the new world of work. HR needs to accept that behavior will be changed most effectively when predictive analytics are used to their full potential. Resist the temptation to do what you know, embrace disruption and recognize where you do not have sufficient skills.
Don’t miss Lisa and Tristan’s conference co-presentation, Predictive Analytics Unlocks Sustainable Cost Reduction In Hourly Workforce, at PAW Workforce, on Tuesday, April 5, 2016 from 3:30 to 4:15 pm. Click here to register for attendance.
The post Wise Practitioner – Predictive Workforce Analytics Interview Series: Lisa Disselkamp and Tristan Aubert at Deloitte appeared first on Analytical Worlds Blog – Predictive Analytics and Text Analytics – by Eric Siegel, Ph.D..
You must log in to post a comment.