Why you should retrain your employees to become your data scientists

Data science has become a hot topic this year, as data-driven campaigns help employees across the board perform better. From marketers to salespeople to consultants, data can enhance performance and help overall company productivity. Businesses agree that data can help, but one issue starting to arise is a talent gap when it comes to data scientists.

International Data Corporation (IDC) predicts that by 2018, businesses in the US will need 181,000 people with deep analytical skills and 5x that number of people with data management and interpretation skills. As of now, however, there are not enough skilled data scientists to meet that need. So how can companies find candidates that can fill these positions? To start, companies need to look internally and begin to realize the benefits of retraining their own staff to become data scientists.

Retraining to give employees new skills

Employees appreciate learning and development. Seven out of 10 employees say that training and development opportunities impact their decision to stay at a company. And 40% of employees who are given poor job training leave their positions within the first year. Without professional growth, an employee is more likely to be unhappy and leave, which means training and development must be a top priority.

A few years back we saw a “coding craze”, where many companies trained their employees to become basic programmers. Companies had everyone, top to bottom, learn to code in order to enhance technical skills across the board and increase productivity. One company, FreeCause, performed a “Codinization Project” that brought together technical and non-technical employees to have everyone in the company learn to code.

It’s safe to say this idea worked and got hundreds of employees across all types of industries to learn to code. Now, what if the same strategy was applied to address the widening talent gap in data science?

Business incentives of retraining employees

Wharton management professor Peter Cappelli took a look at why some companies hire new employees and while others retrain existing employees. Cappelli looked at the impact each had on social capital, or relationships within a workplace and how well employees work together. The results were interesting:

If a firm chooses not to retrain, it replaces existing employees with new ones. In the process social networks in the workplace are disrupted, and social capital is destroyed. If a firm does retrain, it preserves social networks and retains social capital.

The study notes that since many companies rely on social capital to function properly, retraining employees is important.

Financial incentives can be a huge reason for companies to retrain their employees instead of hiring new candidates. Studies show that the direct cost of replacing an employee can reach as high as 50-60% of their annual salary. When you add that to associated costs (delays in production, disruptions to workflow, etc.), the total of hiring a new person can reach 90-200% of their annual salary.

One company that did this was IBM in 2014. IBM decided to retrain employees instead of having layoffs, giving employees 90% of their salary while they learned new skills such as the cloud, analytics, and mobile. Instead of simply firing employees and hiring new candidates with this set of job skills, IBM decided to invest internally.

Retraining employees as data scientists

McKinsey estimates that by 2018, the number of data science jobs in the US will be over 490,000. But the eligible, qualified candidates to fill these jobs will be less than 200,000. On a global scale, the demand will exceed the supply by 50%, according to the report. One issue is that universities are not making undergraduate programs in data science, and those that do cannot keep up with new and changing technologies.

New, innovative data science boot camps and custom corporate training programs allow companies to train existing employees with technical backgrounds to be data scientists in a matter of weeks.  Any employee with a coding background can now be trained to become an expert in R, Hadoop, Spark, and Python, or any custom training curriculum to meet a company’s data science needs. 

With these options, organizations can now consider retraining existing employees rather than hiring new people and incurring social and financial costs. In addition to financial benefits, companies that reinvest in their employees have lower turnover rates and happier employees. So let’s take the lessons learned from successful coding boot camps and apply those to teach employees data science as well.