No More Excuses—Just Get Predictive!
In this final blog post based on our recent interview, Greta Roberts, co-founder and CEO of Talent Analytics, Corp. explains what predictive workforce analytics maturity really means and provides insights for organizations transitioning to making data-based workforce decisions.
Greta will be a keynote speaker at The Predictive Analytics World for Workforce conference, which is being held in San Francisco from April 3rd through 6th in 2016. The conference is the premier workforce analytics event for HR professionals, business leaders, line of business managers, and analytics practitioners. This global, cross-industry event covers predictive solutions to today's greatest workforce challenges. Join Greta Roberts and APQC when you register today with 15% off code APQC15.
Elissa Tucker: One of the conference sessions is going to talk about how the healthcare company, F. Hoffman La-Roche, reached predictive workforce analytics maturity. What does predictive workforce analytics maturity mean to you, and why do you think it’s important for organizations?
Greta Roberts: My answer may be a little unexpected, but to me, the concept of predictive analytics maturity means doing something. It’s having the maturity to do a predictive project versus the maturity to just document your process. Maturity is actually doing something.
We see a lot of organizations that are just learning which is great. But this doesn’t mean they’re analytically mature. They spend all this time documenting, learning, hiring consultants, and then they stall and do nothing. They don’t move forward after having spent a huge amount of time and investment investigating and in some cases even hiring data scientists. At the end of the day, really, who cares if you investigate it, write things down, put it in a pretty report, hire a data scientist, have a list, and know where you are in the analytics maturity path, if you don’t act?
It is important that organizations do not wait until they reach maturity to do something. A lot of people say: “I have to reach analytics maturity, and when I’ve finally arrived, then, I’m going to be able to do predictive projects.” It doesn’t work that way. To me, acting is being the more predictively mature. Do something with the knowledge that you have. That’s what is special about F. Hoffman La-Roche in their journey. They did just that. They went on a journey. They did a lot of discovery. They learned a lot and they did a predictive project which is yielding for them. Is it going to be the same yield this year as next year or 20 years from now when they are even more analytically mature? No, but they did something with the process.
It is important to realize that predictive analytics maturity is a journey and that you’re never going to arrive. It’s something you’re going to always aspire to and to continue to get better and better at. It’s a long-term goal and process.
Elissa Tucker: At APQC we are currently researching how organizations can transition from making business decisions based on intuition, to having an organizational culture of making more data-driven decisions. You work with a lot of organizations that are in various stages of such a transition. What do you think are some of the key steps for an organization trying to move toward making more data-based decisions?
Greta Roberts: The first step, and I’ll sound like a broken record, is don’t feel a need to complete the analytics maturity model before you move into predictive work. Every single organization that we work with that is doing predictive work is doing it right away, even if their data isn’t all in a big data warehouse or data lake. Even if their data isn’t perfect or we need to get data from them in 12 different spreadsheets, we’re doing predictive work that is yielding tremendous ROI. The first step is just realizing the long-term goal is to move toward an even better data model, but concurrently with that, you can begin doing predictive projects.
It’s also important to realize that a lot of the most important workforce data does not exist in HR. HR contains a lot of great activity data such as, start and end dates, employee salaries, other demographics, that sort of thing. But a lot of the most important workforce data exists in the line of business.
For example, if you’re trying to predict who is likely to be a top sales performer, you need to use actual sales performance data. That data doesn’t exist in HR. It exists in sales. If you’re looking to predict bank tellers that don’t make cash drawer errors, that data actually doesn’t exist in HR either. It exists in retail banking. If you’re looking to predict customer service representatives that can process a large number of phone calls, that data exists in call center operations. Yes, there might be some data that exists in HR but again, if we’re thinking of this as an HR problem, we’re going to think of this as HR data. If we think of this as a workforce problem, we’re going to think of workforce data. You need to realize that a lot of the most important workforce data does not exist in HR.
It’s also helpful to start predictive work in positive and benign areas. When we talk to customers sometimes they will say: “We can’t do that.” They’ll find one little area where predictive is not going to work. For example, where predictive pushes all kinds of privacy buttons. They say, “We can’t go there.” At Talent Analytics, Corp. we suggest they start predictive in very positive and benign areas, things that don’t push those privacy buttons or lots of concerns. Go benign. It doesn’t sound exciting, but go into those areas that are going to have a large impact and they’re less scary.
In addition, people need to be prepared to let go of their egos. I may have been at the organization for ten years and was always able to make decisions with my intuition. But now I need to recognize that the predictive model is able to process many more factors and variables than my brain. So if the predictive model comes back with a different result than I would have expected, I need to be okay with that. I need to say: “I was doing the best that I could do inside of my brain, but this model just outpaces me, and it’s okay. It doesn’t mean I was bad or wrong. We’re just doing something new here.”
Finally, people need to be prepared for a different level of accountability. Predictive models can identify what we call mystery factors. They can provide insight into what went into a particular decision. For example, why did I decide to hire one person versus somebody else? I can say, it was their education, their reference checks, or things like that, but there are a lot of other mystery factors behind the scenes that I don’t even know. Predictive models begin to identify these mystery factors and make us accountable. This real accountability of everything being written down and documented automatically is kind of a new thing that people need to be prepared for.
- APQC members can check out our The Path to Predictive Workforce Analytics Success (Collection)
- Non-members can take a look at Predictive Workforce Analytics: Do's and Don'ts (Infographic)