Churn in Data Science Teams – Why and how to prevent it?

No alt text provided for this image

To those in the industry, it won’t come as a surprise to hear there is a lot of churn in data science functions. In this blog, I look at why and explore what the potential solutions could be to prevent it.

Case Scenario:

HR Manager asks the Chief Data Scientist, “Why have you lost 6 people in 3 months?” (This is assuming the chief data scientist hasn’t left already!
Chief Data Scientist replies, “Well frankly our systems and data are all over the place, and no one has a budget to fix it. We don’t have access to any external datasets and our salaries are at least 50% below market. Oh and since you’ve brought this up, I am resigning.”

A lot of leaders are familiar with this conversation.

Data science is a relatively new industry compared to more traditional professions such as accounting, law, marketing, and sales. To make matters worse Data Science has been plagued by a lot of “snake-oil” salesman, all of whom are either trying to sell you software or trying to land themselves a 7 figures job, with none or little background in statistics and data analysis.

For Data Science to be utilized to its full capacity it is dependent on lots of company functions, it’s not just systems, architecture and engineering, but also the business stakeholder function, all of which would apply the insights or products generated by the data science function.

Integrating data science functions into a company can be a struggle and an organizational design challenge. I believe this is something CEOs/CFOs want to do because they can see the value in it, but it’s something that these leaders and their HR departments have struggled to keep up with. As an example, I recently filled a board level Chief Scientist mandate, the chief data scientist came in with a mandate to architect, and build out their data systems and infrastructure; implement strict data governance processes and build out an analytics function. This had all been pre-agreed at interview stage, and there was a clear mandate and budget in place! However, unfortunately, this is not usually the case, and data scientists end up spending most of their time sorting out systems and data issues, rather than actually building predictive/prescriptive models, where the real ROI and fun is.

We can learn a lot from case studies, Eric Colson at Stich Fix is a great example of someone who has built an amazing data science function and culture, but he has been empowered to do so. As far as I know he sits at board level and reports directly to the CEO, data science is not ‘a supplement’ or the ‘red-headed stepchild’ (no offense to red heads) of Marketing, Sales, Risk, or other business functions that are more established in the company but seen as a game changer.

Whilst there is no easy or cheap solution, your organization should start by empowering your data science team. Give them the right tools and data, and don’t silo them across multiple departments and management groups, as this creates a lot of room for resentment, politics, and little room for getting good work done.

However, if you feel like you’re in this situation now, connect with me on LinkedIn for a confidential chat or email me at [email protected]

I only work with employers who have already established a serious data science function and have already taken steps (often by learning from past mistakes) to ensure that you will be successful in your role.  


要查看或添加评论,请登录

社区洞察

其他会员也浏览了