You’ve scaled your business to the point where you have product-market fit. Your leadership team has good intuition about the market, product, and opportunities in front of you. Yet the opportunities are too vast. You’re unable to assimilate data required to understand tradeoffs, measure ROI, and make decisions that feel well informed.
Perhaps it’s time to hire your first data scientist.
I gave a talk recently about building a data-driven organization recently (see Tabi’s overview of my talk on the Simon Data Blog), and I’ve been getting a ton of questions on how to hire a data scientist. While the data science job may be one of the sexiest & fastest growing jobs on the market, it’s also one of the most poorly defined and covers a diverse skillset.
How should you think about making this first hire? Below is four critical qualities for evaluation.
Do they understand the business? Data science ends with crunching numbers but starts with having a deep understanding of the context in which these numbers live. Good science is driven by strong hypotheses, and strong hypotheses require understanding the problem you’re solving. Your first data science hire should naturally orient to problems that your business faces. Do they understand inherent challenges with your customer acquisition strategy? How are you thinking about current customers and early adopters compared to your next tranche of customers? It’s critical that your data scientist understands core business considerations and tradeoffs.
Can they communicate? Data should facilitate conversations and serve as an equalizer across the organization. Hiring your first data scientist marks a transition in decision making. Data shouldn’t be the deciding factor but instead a core consideration; trade-offs around assumptions, intuition, and strategic bets must be understood. The role of data science needs to be one of facilitating a conversation around data and trying to quantify such trade-offs when possible.
Are they analytical? Your first data scientist should have experience in modeling business processes. Give them a problem - modeling customer lifetime value, customer acquisition costs, supply chain issues, etc - and they should be able to identify assumptions & inputs, build a model, and explain the outputs. They should have an insatiable desire to improve this model and map it back to available data.
Do they know SQL? Your first data scientist should be technical but not overly-so. It’s more important that they know the basics; SQL proficiency and a mastery of Excel is a must, knowledge of Python or other numerical computing languages is a plus. As your company grows, your data science requirements will become more sophisticated - and assuming that this hire is intelligent and self-motivated, they’ll pick up more advanced skills on their own.
There’s a continuum between a data analyst, a data scientist, and an AI researcher. Arguably your first data scientist could just be an analyst - and arguably the qualities above are relevant to an analyst role. Yet focusing on data science provides more upside - predictive insights & AI are rapidly maturing. Just yesterday, Amazon made their internal machine learning courses free for anyone and given the rapid growth of the field, it’s hard not to start investing in a data scientist today.