3 things to consider before hiring a Data Scientist

It is commonly accepted that data science can bring tremendous value to an organization. That being said, a pitfall for companies when pursuing data science initiatives is hiring data scientists without having a clear vision of their goals, business impact, and expected results.

Before embarking on the long (and expensive) journey of hiring a data scientist, take a step back and make sure your organization is data science ready. This includes developing a concrete, results-oriented data science strategy and auditing your underlying data to ensure your data is accurate, consistent, and complete enough to enable reliable analysis.

Step 1: Develop your data science strategy

The process of hiring a data scientist requires a tremendous amount of time, money, and effort. It could cost your company up to $30,000 just to find a candidate with the right skills and personality to fit your business. Along with the ever-increasing salary, which currently averages around $113,000 (excluding benefits), this is a huge investment. If you hire a data scientist without having a clearly defined business objective for data science, you run the risk of burning that investment and draining talent.

Showing a candidate you have a strategy will inspire their confidence in your organization and help them determine if they are up to the challenge. If you’re considering hiring and onboarding a data scientist, you shouldn’t leave it up to them to determine their mission and place. To start developing a strategy, ask your IT team and business leaders to join forces to find answers to some of the questions below:

  • What are our business problems and opportunities? Do the goals of our data science initiatives align with the goals of our organization?
  • What data do we have to support the analysis?
  • What business or metric definitions vary across our organization? Why do these knowledge silos exist and how can they be overcome?
  • Can our current infrastructure meet data science needs?
  • Are we ready to change as an organization based on data science initiatives?
  • How can we effectively communicate data science results?

Step 2: Assess your company’s data science readiness

Accurate and readily available data is essential for any data science project. The quality of the data you use for analysis has a direct impact on your result. In other words, if no one trusts your results, they won’t use that information to inform their decision-making, and your entire data science strategy will fail. Prepare your data science team for success by providing clear, centralized data so they can be up and running.

While your data doesn’t need to be perfect, you should at least make sure your data is centralized and doesn’t contain duplicate records or large amounts of missing information. Centralizing key information in a data warehouse eliminates time wasted searching for data or finding ways to circumvent data silos. Creating a system that cleans, organizes and normalizes your data ensures reliable information for everyone. This will not only help your new data scientist produce results faster, but it will also increase confidence in their results around your organization and save hours of menial data cleaning done by your IT team. Although the steps to prepare for data science are different for each company, they should all take into account the same objectives.

Step 3: Define clear and actionable business cases for data science

An important part of a successful data science strategy is understanding the insights data science can provide and how your business can act on that insight. Start by thinking about a variety of use cases. Determine which are the most actionable, relevant and offer the best competitive advantage. If one of your ideas can save you money, that’s another great place to start. During this process, there are no wrong answers. Identifying use cases can seem daunting at first, but there are some very simple ways to get started:

  • Ask employees what common business questions go unanswered.
  • Examine what industry leaders (and your competitors) are doing. From personalizing marketing messages to customers to using patterns to identify insurance fraud, data science has use cases in any industry.
  • Find out what leaders would like to be able to predict about your organization.
  • Call in the experts. Many organizations (consulting firms and vendors) have implemented data science solutions at various clients.
  • Identify time-consuming and complicated manual processes. Data scientists can probably automate them and make them more reliable.

Ryan Lewis is a management consultant at 2nd watch.

Sean N. Ayres