Data scientists and decision scientists do very different, though equally important, work. Here’s how to tell the difference.
At Instagram, we had many different job roles that analyzed data. A few of the data job titles included: data scientist, analyst, researcher and growth marketing.
There’s often a lot of confusion between the roles of data scientist vs. decision scientist.
We had both at Instagram and they fulfilled different needs, so I thought I’d explain the main differences I see from my personal experience in the decision science role, working closely with my data science colleagues.
DATA SCIENCE VS. DECISION SCIENCE
The data scientist focuses on finding insights and relationships via statistics. The decision scientist is looking to find insights as they relate to the decision at-hand. Example decisions might include: age groups on which to focus, the most optimal way to spend a yearly budget or how to measure a non-traditional media mix. For decision scientists, the business problem comes first; analysis follows and is dependent on the question or business decision that needs to be made.
DATA SCIENTISTS
The data scientist focuses on finding insights and relationships via statistics. The decision scientist is looking to find insights as they relate to the decision at-hand. Example decisions might include: age groups on which to focus, the most optimal way to spend a yearly budget or how to measure a non-traditional media mix. For decision scientists, the business problem comes first; analysis follows and is dependent on the question or business decision that needs to be made.
DATA SCIENTISTS
Data is the Tool for Improving and Developing New Products Based on Robust Statistical Methods
Data scientists are looking to understand, interpret and analyze with the goal of building better products. Therefore, data quality, statistical rigor and measurement perfection are often their trademarks.
For data scientists, the analysis, statistical rigor and understanding comes first. Business challenges come second. Data scientists think about data in terms of data patterns, data processing, algorithms and statistics. Often, data scientists are conducting deep analysis and experimental statistics. They are obsessed with finding causal relationships.
Data scientists are deeply focused on data quality as it relates to their product area because better data quality results in more thorough statistical analysis.
Data scientists frame data analysis in terms of algorithms, machine learning, statistics and experimentation. They are looking to bring order to big data to find insights and learnings as they relate to their product or focus area. Theyhave a statistics lens to everything they do.
Data scientists’ north star goal: Use high-quality data and robust statistics to support product development.
DECISION SCIENTISTS
Data scientists are looking to understand, interpret and analyze with the goal of building better products. Therefore, data quality, statistical rigor and measurement perfection are often their trademarks.
For data scientists, the analysis, statistical rigor and understanding comes first. Business challenges come second. Data scientists think about data in terms of data patterns, data processing, algorithms and statistics. Often, data scientists are conducting deep analysis and experimental statistics. They are obsessed with finding causal relationships.
Data scientists are deeply focused on data quality as it relates to their product area because better data quality results in more thorough statistical analysis.
Data scientists frame data analysis in terms of algorithms, machine learning, statistics and experimentation. They are looking to bring order to big data to find insights and learnings as they relate to their product or focus area. Theyhave a statistics lens to everything they do.
Data scientists’ north star goal: Use high-quality data and robust statistics to support product development.
DECISION SCIENTISTS
Data is the Tool to Make Decisions
Decision scientists frame data analysis in terms of the decision-making process. They are looking at the various ways of analyzing data as it relates to a specific business question posed by their stakeholder(s).
Other names for this role may include: analytics, analyst and applied analytics.
The data scientist focuses on finding insights and relationships via statistics. The decision scientist is looking to find insights as they relate to the decision at-hand. Example decisions might include: age groups on which to focus, the most optimal way to spend a yearly budget or how to measure a non-traditional media mix. For decision scientists, the business problem comes first; analysis follows and is dependent on the question or business decision that needs to be made.
The decision scientist therefore needs to take a 360-degree view of the business challenge. They need to consider the type of analysis, visualization methods and behavioral understanding that can help a stakeholder make a specific decision.
In other words, decision scientists need to make insights usable. They need to be able to work with a variety of data sources and inputs — each selected based on its ability to help answer the business question. This means a decision scientist needs to have a strong business acumen as well as a robust analytical mind. You cannot have one without the other in a decision science role.
Sometimes, measurement won’t be perfect. Business tactics aren’t always neat and tidy. For example, there is almost no clean way to create a test and control for viral or celebrity marketing, but these are both legitimate marketing approaches and the decision scientist needs to be okay with that. Businesses shouldn’t take an action so that it can be measured, but because it is the right thing to do; measurement comes next.
Sometimes a clean, causal experiment is possible and sometimes it isn’t. Decision scientists need to have a keen sense of when it’s appropriate to move forward with a decision based on correlations and when they need to push for a clean experiment. It all comes back to the business context and the decision at-hand.
Decision scientists’ north star goal: use data and statistics to support business decision making, budgeting and marketing spend.
Why Decision Science Matters
Data Science vs. Decision Science: In the Real World
In my own experience at Instagram, each data scientist was dedicated to one specific product or product feature. They spend a lot of time ensuring the data logging is accurate for that product area by running statistical analysis on trends and using complex visuals to display their type of analysis. They have a deep knowledge of their product, but not the ecosystem.
If the product changes or we launch new features attached to their product, the data scientist is responsible for both logging the new data and measuring the uptake of the new features.
On the flip side, I was in the decision science job group. My team and I supported the marketing group and the marketing leadership in helping them make decisions about marketing budgets and priorities.
I relied heavily on the tables, logging and analysis from my data science colleagues as the basis for our marketing activities. I then augmented their work with my own analysis to help our marketing leadership make decisions on where and when to spend marketing budget.
My visuals were designed for consumption and business action, and therefore had a different goal than the data scientists’ goal of using visuals to display complex analysis.
Because data scientists focus on one product area only, my analysis tended to look at relationships across products and the impact of demographics on product behavior at the company level.
My decision science team is the only team that looks at the full ecosystem on a regular basis because marketing decisions revolve around wanting to understand how one behavior interacts with another.
As you can hopefully see, there are some subtle but important differences here.
The decision scientist sits hip-to-hip with decision makers and management to help them make the best decisions for the business. Decision scientists are equal parts business leader and data analyst.
The data scientist sits hip-to-hip with data and statistical rigor. Data scientists are relentless about quality and deep analyses that drive products to scale and develop based on usage data.
Each role is necessary and critically important.
Decisions need to be made quickly to keep the business moving forward based on what is knowable now. This is the job of the decision scientist.
The business also needs to grow, scale and build better products. Deep product knowledge, a high standard of data quality and statistical rigor help ensure they’re pulling out the best insights so product leaders understand their domains. This is the job of the data scientist.
A business needs to both move forward with decision making while also improving its products for the longer term, so the decision scientist and the data scientist both contribute to the greater health of the company.
Decision scientists frame data analysis in terms of the decision-making process. They are looking at the various ways of analyzing data as it relates to a specific business question posed by their stakeholder(s).
Other names for this role may include: analytics, analyst and applied analytics.
The data scientist focuses on finding insights and relationships via statistics. The decision scientist is looking to find insights as they relate to the decision at-hand. Example decisions might include: age groups on which to focus, the most optimal way to spend a yearly budget or how to measure a non-traditional media mix. For decision scientists, the business problem comes first; analysis follows and is dependent on the question or business decision that needs to be made.
The decision scientist therefore needs to take a 360-degree view of the business challenge. They need to consider the type of analysis, visualization methods and behavioral understanding that can help a stakeholder make a specific decision.
In other words, decision scientists need to make insights usable. They need to be able to work with a variety of data sources and inputs — each selected based on its ability to help answer the business question. This means a decision scientist needs to have a strong business acumen as well as a robust analytical mind. You cannot have one without the other in a decision science role.
Sometimes, measurement won’t be perfect. Business tactics aren’t always neat and tidy. For example, there is almost no clean way to create a test and control for viral or celebrity marketing, but these are both legitimate marketing approaches and the decision scientist needs to be okay with that. Businesses shouldn’t take an action so that it can be measured, but because it is the right thing to do; measurement comes next.
Sometimes a clean, causal experiment is possible and sometimes it isn’t. Decision scientists need to have a keen sense of when it’s appropriate to move forward with a decision based on correlations and when they need to push for a clean experiment. It all comes back to the business context and the decision at-hand.
Decision scientists’ north star goal: use data and statistics to support business decision making, budgeting and marketing spend.
Why Decision Science Matters
Data Science vs. Decision Science: In the Real World
In my own experience at Instagram, each data scientist was dedicated to one specific product or product feature. They spend a lot of time ensuring the data logging is accurate for that product area by running statistical analysis on trends and using complex visuals to display their type of analysis. They have a deep knowledge of their product, but not the ecosystem.
If the product changes or we launch new features attached to their product, the data scientist is responsible for both logging the new data and measuring the uptake of the new features.
On the flip side, I was in the decision science job group. My team and I supported the marketing group and the marketing leadership in helping them make decisions about marketing budgets and priorities.
I relied heavily on the tables, logging and analysis from my data science colleagues as the basis for our marketing activities. I then augmented their work with my own analysis to help our marketing leadership make decisions on where and when to spend marketing budget.
My visuals were designed for consumption and business action, and therefore had a different goal than the data scientists’ goal of using visuals to display complex analysis.
Because data scientists focus on one product area only, my analysis tended to look at relationships across products and the impact of demographics on product behavior at the company level.
My decision science team is the only team that looks at the full ecosystem on a regular basis because marketing decisions revolve around wanting to understand how one behavior interacts with another.
As you can hopefully see, there are some subtle but important differences here.
The decision scientist sits hip-to-hip with decision makers and management to help them make the best decisions for the business. Decision scientists are equal parts business leader and data analyst.
The data scientist sits hip-to-hip with data and statistical rigor. Data scientists are relentless about quality and deep analyses that drive products to scale and develop based on usage data.
Each role is necessary and critically important.
Decisions need to be made quickly to keep the business moving forward based on what is knowable now. This is the job of the decision scientist.
The business also needs to grow, scale and build better products. Deep product knowledge, a high standard of data quality and statistical rigor help ensure they’re pulling out the best insights so product leaders understand their domains. This is the job of the data scientist.
A business needs to both move forward with decision making while also improving its products for the longer term, so the decision scientist and the data scientist both contribute to the greater health of the company.
www.dprg.com
