Talent Analytics CEO Greta Roberts (@GretaRoberts) will be presenting at SAS’s Analytics 2013 Conference next month in Orlando, Florida on the results of the landmark 2012 Analytics Professionals Study.
I recently had the opportunity to get Greta’s thoughts on a myriad of topics including what to expect at Analytics 2013, core skills for analytics jobs, and what the future holds for analytics.
What are the hurdles to using analytics and big data effectively?
GRETA: It’s such a great question. Two of the greatest hurdles are 1) an adequate supply of analytics professionals to solve complex business problems, and 2) adequate demand of strategic projects for analytics professionals to work on that will yield high value outcomes.
I’m beginning to believe (and see data to support) the biggest barrier to using analytics and big data effectively might be a lack of demand of important, strategic analytics projects to work on.
The world and many consultants might have it backwards.
Many businesses believe analytic talent is responsible for knowing which problems to solve and which projects to do – this isn’t the case. Business leaders are the ones who need to identify the problems and then ask analytics professionals to solve them.
This goes against conventional wisdom that analytics has a talent supply problem, an area I addressed in a recent INFORMS podcast and will discuss in my presentation at Analytics 2013.
What are some best practices for managing big data to improve decision making?
GRETA: For big data best practices, I recommend picking up a copy of "Taming the Big Data Tidal Wave" by Bill Franks the CAO at Teradata and a colleague of both ours and SAS’s as well. Bill's book is a very manager friendly read and loaded with best practices around managing big data.
That being said, improved decision making using analytics typically refers to improved decision making about customer decisions. As it turns out – customers are people. Employees are also people.
Our approach has been to work with customers to use an analytics approach to improving decisions with another set of people – employees. In the same way businesses calculate ratios of 1 bad credit decision erases 4 good credit decisions – we work with customers to learn ratios with employee decisions. One example: does 1 bad hiring decision erase 4 good hiring decisions?
We are taking advanced analytics used with customers and evolving this into better (predictive) decision making that optimize employee performance.
What do you think will be the big topic that has everyone talking at Analytics 2013?
GRETA: I may be biased but talent is central to any successful analytics initiative. Just as it has been covered in the media, I expect attracting and retaining data scientists to remain a hot topic at Analytics 2013.
Specifically, I believe there will be many conversations around reconciling whether the oft-cited analytics talent supply problem is actually an analytics project demand issue, reflecting a shortage of analytics projects inside organization. This concept resonated with my analytics practitioner audience at SAS Germany in September and as mentioned before, I will address this topic in my presentation.
What advice do you have for attendees of Analytics 2013?
GRETA: The Analytics 2013 agenda brings together the top minds in the field, offering unprecedented networking and learning opportunities for attendees – take advantage of them. For example, on Tuesday morning my track includes a phenomenal panel discussion of top academics representing leading analytics programs, facilitated by our colleague and genuine rock star Dr. Jennifer Priestley. If analytic talent is strategic to your firm’s success, I would encourage you to attend this panel as well as my session on 10/22 at 3pm.
What advice would you give students considering a job in analytics and those already in the workforce to put themselves in a position to pursue these jobs?
GRETA: I would urge folks considering the field to be honest with themselves about why they might be pursuing this. If it is because they genuinely thirst to solve complex problems and find themselves constantly researching and learning and creating interesting experiments then I would urge them to use their curiosity to research how others get into roles, what works best / doesn't.
Another suggestion is to – on their own – tune their skills. Like it or not, analytical / software skills are one huge metric employers are looking for. I would suggest these folks learn on their own, attend Meetup Groups where they can learn from others, and create their own experiments they can use as examples when interviewing.
The last suggestion is to read a recent blog we published from a current, young Data Scientist. This blog may also provide some guidance to those considering the field – The Musings of a Young Data Scientist. This provides an inside view directly from a successful Data Scientist in the field.
What skills or training would you consider core to any big data job?
GRETA: I personally interviewed 12 leading Analytics Thought Leaders and hiring executives a few months ago. I asked them: “What is the #1 mistake people make when hiring data scientists.” Each of them had the same answer: they all said “hiring for skills is the #1 mistake when hiring data scientists.”
Each person said the most important characteristic is “mindset” not “skillset.” This was consistent with a joint Study we conducted with the International Institute for Analytics. We found that “deep dive” analytics professionals had (among others) a very strong fingerprint for 1) Curiosity 2) Creativity and 3) having a structured approach to problem solving.
We call these characteristics “mindset” or “raw talent” and have heard over and over again from employers that these are more critical than skills for any big data job. It makes sense; technology is changing so fast and employers are not necessarily looking for what you’ve learned but what your capacity for learning is.
What’s the biggest change you see in the analytics field when you look out for the next 15 years?
GRETA: Fifteen years from now analytics will have proliferated into predicting and optimizing employee performance. While this may seem perhaps “too close” today, I have no doubt that even 2 years from now employee performance optimization will be commonplace. I predict that employers will attach a financial value to “their most important asset” as they are able to attach a financial value to all other assets.
Additionally, fifteen years from now, “black box analytics models” will be commonplace and data scientists will have proven their value.