Keeping Data Scientists Through Tech Acquisitions: M&A Talent | Deloitte US has been added to your bookmarks.
How do you retain your data scientists?
Keeping tech talent through a post-merger integration
Retaining talent after a merger or acquisition is a common challenge for leadership. Especially as more companies scramble to add artificial intelligence and cognitive computing to their core competencies, what steps can they take to keep data scientists happy during post-merger integrations?
- The demand for data scientists
- What do employees want?
- Office politics
- What drives job satisfaction of Data Scientists?
- Learning opportunities
The demand for data scientists
Data scientists are in demand. And they know it. They spend an average of one to two hours a week looking for new jobs and welcome offers from potential employers. As noted in a recent Deloitte report, these new employees are often millennials who are passionate about creating technology that could change the world.
With data scientists in short supply, it’s critical to retain them following a merger or acquisition. But their new company will likely differ from their previous firm: Bigger, more process-driven, and less focused on the “mission” they signed up for. And they may resent that the start-up entrepreneurs who originally hired them—who may have been classmates or friends—might play greatly diminished roles in the new organization.
What do employees want?
Using sophisticated machine learning algorithms on a global survey of more than 4,000 data scientists conducted by Kaggle, Deloitte gained unique insights into the mind-sets of these frontline artificial intelligence (AI) workers and how changes in their work environment could affect their job satisfaction.
Using a 10-point scale of job satisfaction, with a higher score associated with a happier employee (the average was 6.9 in the subset we studied), we looked to see which factors had the most substantial predicted impact on job satisfaction.
The single largest effect we observed involved office politics, which can be a serious problem for data scientists because many feel poorly equipped to handle it. And companies that are building data science teams may struggle to provide the support and direction they need—especially if they’re new to the game. Data scientists in a strife-ridden work environment—compared with one free of infighting, and with all other factors being equal—had job satisfaction that was 1.3 points lower, making it the biggest move we saw in the entire data set.
But it would be a mistake to think that allowing them to work remotely would be an effective bandage for a difficult office environment. We discovered that the more people work off-site, the more affected they are by political issues: Remote workers in politicized work environments experienced a job satisfaction decline of 1.5 points, compared with a decline of 1.2 points for employees always in the office. Clearly, a strong corporate culture gives you the flexibility to allow more remote work. However, remote work can exacerbate underlying issues if your corporate culture is fractious.
Our analysis also uncovered the importance of training: Survey respondents showed they expected to learn "on the job" by being exposed to new tools and techniques. Companies that have an internal learning platform show an anticipated 0.3-point improvement in job satisfaction compared with those that don’t. So if your programmers are excited about a hot new analysis package, get it in their hands.
Data scientists like using new technologies that help them build on their experience. But they don’t like being thrown in the "deep end" of subject matter outside their domain and then being expected to swim. This often happens when programmers and business leaders don’t communicate effectively about project goals and expectations.
Strong project management and "translators" between the business and data scientists can ensure that expectations on both sides are clear and reasonable—and that the data scientists feel they know how to do what’s required.
Not surprisingly, when we looked at compensation, we discovered that may levels are important. We were intrigued to learn, however, that compensation momentum is even more important than the size of an employee’s current paycheck.
Data scientists are more likely to accept relatively modest salaries if they know their pay increases point to opportunities in the future. And sometimes, simply recognizing high achievement is all it takes. Changing a job title to make it more representative of an individual’s actual duties resulted in a predicted 0.5-point increase in job satisfaction.
Visit our "M&A in Tech, Media & Telecom: Charting a well-defined integration strategy" pageRead the report
Start with the end in mind
Part of the User Friendly podcast series