The Best Data Scientist is not one person, it’s an Interdisciplinary Team
One of the most common mistakes organizations make when they start to get serious about using data is to hire a data scientist. Why?
Because “data science” is a function, not a person. To do data science right, you need a team of people with diverse hard and soft skills.
Who is on the Team?
Business Thinker (a.k.a. Business Analyst)
The Business Thinker understands the business cases, questions, problems and opportunities that are facing a company or organization. They understand B2B or consumers that are organizations and they understand B2C or consumers that are individual purchasing decision makers. In addition, they have strong knowledge about specific or multiple industries, their competition, the marketplace and the external factors influencing markets. In addition the Business thinker oversees data development, analysis, and application processes to make sure the team is operating cost-effectively. They can spot industry trends that could impact the organization. And they ensure that the end-user gets what they need. What makes the Business Thinker especially valuable is their in-depth knowledge of the organization and its industry and have a good understanding of data, without being an expert. This lets them see insights in the data that helps your business move in the right direction. A Business Thinker works closely with the Data Thinker to translate business cases into data and data into clear business explanations.
Data Thinker (a.k.a. Statistician, Mathematician)
The Data Thinker once they understand the business case and questions works to understand what data is available currently, and what data needs to be collected to answer the specific questions. A data thinker must be able to understand the data, know whether it is good data (clean, normalized, regularized and imputed), determine what analysis needs to be performed, and be able to explain the results and offer explanations on how to verify the results. A Data Thinker keeps your team running by using statistics and math to collect, analyze, and interpret the data you’ve collected. They also help determine the method to use when collecting data for a specific purpose. Data Thinkers are especially valuable when your company collects data through techniques like surveys, field experiments, and focus groups. In addition, to implement Machine Learning, a data thinker needs statistics, linear algebra and calculus. The Data Thinker is assisted and directed by the Business/Consumer Behavior thinkers.
Consumer Behavior Thinker (a.k.a. Psychologist)
The Consumer Behavior Thinker is often a psychologist or behavioral scientists that understands human motivation, human behavior and psychometrics. How you ask questions and enter data can be a key deciding factor in whether your data science results are accurate. A Consumer Behavior Thinker helps companies ask the right questions, input data and rating scales correctly, understands statistics and behavioral data, and helps companies make sense of consumer data. This individual must be able to tell stories with the data that C-Level Executives understand. If the data science answers don’t make sense to anyone, you are left with an expensive opinion. Their ability to interpret numbers and human behavior makes them key members of the team. The Consumer Behavior Thinker can identify variables relevant to the business question you are trying to solve with your data and make sure you are asking the questions and entering the responses correctly.
Data Jock (a.k.a. Data Algorithm Programmer)
The Data Jock is a data software engineer that can design and architect the collection and management of a companies data, work with the Data Scientists to ensure the data is clean and normalized, and most importantly help to build the software algorithms and data extrapolations from the data science results. The Data Jock also has an in-depth knowledge of industry-leading analytics and data frameworks and can apply them, as well as the hosting, importing and exporting of data among servers. They’ll know programming languages, such as SQL, R, and Python. They can create complex analytical models that make the most of the gathered data to give your organization the information it needs. The Data Jock works closely with and receives direction from business/consumer behavior/data thinkers to ensure their algorithms and extrapolations make sense and address business issues!
How Horizon Can Help
Horizon Consumer Science has an interdisciplinary team with significant experience in the field to ensure your data science function does what it is supposed to do. And, that we represent much less risk from a time and cost perspective than recruiting and hiring a Data Science team internally.
Often data science needs an ebb and flow, as there will be times of high-intensity working and other times with little activity. That’s one reason why it may be more valuable to outsource. It becomes a variable cost, rather than a fixed cost which may make a lot of sense for your bottom-line.
Another good reason to outsource is the shortage of qualified technical data science team members. You can find a good partner for your ad-hoc or routine data science needs.