“A data scientist does model-driven analyses of our data; analyses to improve our planning, increase our productivity, and develop our deeper levels of subject matter expertise. A data scientist works at the tactical, operational, and strategic levels, sharing insights with the business.”
Fortune companies bring in data scientists, shoe horning them into their organizational structures; structures that don’t support the data scientist or the data-driven organization. This shoe-horning compresses the data scientist to be an IT person that analyzes data, not a data person that analyzes data using IT.
“For data scientists to be successful we must first restructure and align the organizational structure so our data scientists and Big Data experts can grow and thrive. We can’t just shove data scientists into our IT organization and expect it to work.”
To make our data scientists fit into our organization, we must transition our organization to become more data-driven; and we do this by building a data-driven talent pool. Our talent pool includes more than just data scientists. Our talent pool includes data practitioners; people in traditional roles that use data-driven approaches; people that are almost, but not quite, data scientists.
“Look at how our more analytical talent plan, forecast, and make decisions. They use models, they use algorithms, and they use data. If they automated what they did using more volumes of data they would be data scientists. Directors, managers, management scientists, program managers, project managers, business analysts, operational researchers, data analysts; all can be data scientists with the right training.”
This means most of us and most Fortune companies already have the raw talent to develop their own data scientists. What’s stopping us from doing so is our limited understanding of data science, the data scientist, and the data-driven culture. To have successful data scientists, we must first identify our current talent’s fundamental data skills; then build on them to become data scientist skills.
“Our current data practitioners have fundamental data skills. They have a SLAM orientation with a SEAD emphasis. Data practitioners are SLAM oriented with Strategy, Leadership, Analysis, and Management competencies. Data practitioners are SEAD emphasized with Statistics, Engineering, Accounting, and Data algorithms capabilities. By looking at these fundamental skills we can quickly transform our current talent into data scientists.”
HR must bring in the right talent to build a healthy talent pool. HR must bring in the right talent to support our organization’s entire life; how the organization started, grown, acquired, merged, and spun off. Good data scientists don’t just analyze data. Good data scientists build good businesses by analyzing data.
“Our business’s top priorities are having the right customers, the right human capital, and managing our corporate capital the right way through our acquisitions mergers, alignments, and spin-offs. Having the right data is crucial for these key business activities. This is where the data scientist comes in.”
To best support these business activities, we must have our data office promote and advocate our organization’s data capital, planning, innovation, transparency, unification, and reinvention. On top of that, our data office must churn out data scientists by developing and building on our current talent’s fundamental data skills.
“SEAD skills are the fundamental skills for data scientists. For Statistics, we must emphasize the difference between causation and correlation. For Engineering, we must emphasize industrial engineering, models, performance, and productivity. For Accounting, we must emphasize cash flow, returns, investment, and health ratios. For Data algorithms, we must emphasize their correctness, accuracy, and best use. Our data office must build these fundamental data scientist skills through the right hires, the right policies, and the right training.”
Our data scientists must not only have the right training, but also the right career tracks; career tracks where they can advance into management and leadership. This will reinforce and accelerate our organization’s acceptance and integration of data scientists.
“For our data scientists to thrive, we must scaffold over the bottomless pits. Pits include how executives interact with data; how strategic change integrates with culture; how business development integrates with operations; and how business people work with data practitioners and data scientists. These pits are the sources of failure for many; may they be data practitioners, data scientists, or Big Data experts.”
So we can’t put them in there. Instead we must have our data scientists be part of our data office; an office designed to house and produce our top data scientists by developing and promoting our own talent within. Once this is in place, only then will data scientists fit and thrive within our organization.
Thanks Benoit. When introducing Data Scientists it’s always a good idea to map out the data management processes before hand, then use those mappings as a spring board to delegate data responsibilities and accountabilities.
Thanks Chris for this post. I once had a very interesting conversation with an IT consultant & expert and he said that there’s indeed often a disconnect between IT and other departments, including data science teams. He said that IT teams often set their own agenda, and that while they think they are helping employees to be more productive, they often work against them, trying to keep as much control as they can. This is probably why so many new data science teams have a hard time doing their job at the beginning. So I thoroughly agree with you, but in the short term, I guess that management should make sure that IT and new data science teams work closely and cooperatively.