Role of a Data Science Manager

In The Building Blocks of a Data-Informed Company, we discussed what it takes to build a successful data-informed company, namely relentless focus on impact and a data-informed culture. We discussed the role of a data scientist, the different data science archetypes and how to interview for them and the career progression of a data scientist. Even with the best individual contributors in place, a sustainable high achieving organization can also be achieved with great managers. In this post, we’ll dive into the two most important responsibilities of a data science leader: driving impact and building a world class team.

DRIVING IMPACT

Focus on impact is the only real way of scaling a data-informed company. In order to focus on impact, it is necessary to understand and define impact first, and then cultivate the right environment in which to make that impact. This is the job of a data team manager.

Data team managers focus on impact by defining product success and by setting the right goals, metrics and processes to objectively quantify, measure and track impact. Without this, it is very hard for a company to become truly data-informed and achieve its highest potential. Generally speaking, impact can happen when we move a metric and/or influence a product or process change. Ultimately, the litmus test for a good data manager is whether their team has demonstrably improved a product.

Here are a few things that managers should consider as they try to maximize impact.

Set the right metrics and goals: Choose the right metrics, set the right goals, and then be flexible.

• Don’t choose vanity metrics. Choosing the right metric that is measurable is critical for impact. Choosing vanity metrics may make a team feel good, but it won’t create real impact.
• Set the right goals. Impact is achieved by setting the right goals. The right goals are achievable but not easy. Using stretch 50/50 goals (goals that have a 50% chance of being achieved) can often be a good way to push the boundaries of a high-achieving team.
• Be flexible. Products evolve quickly and these days, so do entire industries. Data team managers need to be nimble in adapting their team’s metrics and goals to match.

Define impact: Focus on outcomes rather than inputs, don’t mistake motion for progress, and invest in teams with a high impact per capita.

• Focus on the outcome and not on inputs. The role of a data team manager is to foster a great work environment that creates impact. Interesting work is not impact, nor is an actionable insight. Insight that leads to real change is impact. Managers must hold analysts accountable for influencing eventual positive change that will lead to real impact. Inputs such as working hard matters to the extent that it created a good work culture, but it does not always lead to impact.
• Don’t mistake motion for progress or impact. Motion is not the same as impact, and many people conflate the two. Managers need to keep their team unwaveringly focused on impact and prevent any busy work from landing on their plates.
• Focus on impact per capita — Good data team managers are focused on hiring high-quality talent, setting a high bar of excellence and putting that talent to work on the right goals, rather than building large teams with mediocre talent. Organizational leaders should incentivize creating the greatest impact with the fewest number of people.

Use the right process: Leverage a small number of clear processes and a focus on depth over breadth to maximize impact. While it sounds obvious, having the minimum number of necessary processes can help create the greatest impact. Processes to improve the quality of work, improve teamwork, and share knowledge are all good examples.

• Do a few things very well. Very few things really matter for impact. All things are not equal in terms of impact. Only the top few things that one does really moves the needle. As a result, most times impact can be measured by the total impact of the top few things any individual (or team) has accomplished. The Managers need to ensure that their people are not spread too thin, and that what they are doing drives impact.
• Kill projects with low impact. Managers can boost morale and productivity while freeing up time and resources for higher impact projects by killing low-impact projects, or better yet, not taking them on in the first place.

Invest in the right people: Top talent drives the highest impact. Good managers hire the best people, and put the right people on the right projects.

• A manager needs to identify and recruit top talent, and then put that talent to work solving the toughest problems with the most impact.
• Managers also need to invest in developing their team members so that they can operate at increasingly higher levels of problem solving. Helping people operate at a different level can drive greater impact. For example, taking an analysis from one area and applying it to another can broaden the scope of a team member’s impact.

BUILDING A WORLD CLASS TEAM

In our previous post on building a world-class team, we suggested that three factors — people, culture, and process — are most important to building a strong data team. Here, we’ll take a look at the role a manager in creating a world class team. A manager needs to build a sustainable happy team that is building great products by producing high quality work. Broadly, a data team manager focuses on four things:

1. People management: Does the team have the right people? Are they happy? Are they being mentored and developed?
2. Driving excellence: Is every person on the team maximizing their impact?
3. Scaling the team: How does the team need to evolve over the long term?
4. Product leadership: Are the managers providing the best direction for building products?

People Management

The most important role of a manager is to keep their team members motivated, happy, and focused on high-impact work. Even with the greatest impact and amazing product, a great company can never be built with unhappy employees.

• Keep employees happy. Many companies conduct periodic employee surveys to assess how happy their employees are, including with respect to organizational health, employee welfare, manager effectiveness, and dedication to the company’s mission. This and other types of feedback (regular check-ins, formal reviews) are good indicators of whether team members are happy with their roles, the team, and the company. Feedback mechanisms also give managers an indication of what is working and what is not, and what they need to improve on and advocate on behalf of their team members.
• Actively mentor and coach team members. Managers are ultimately responsible for the welfare and growth of their team members. They need to provide the right mentorship and coaching opportunities to develop the talent they manage. These opportunities need to be tailored to each employee given their interests and growth opportunities.
• Inspiring and motivating the team. Ideally, every employee should come to work every morning feeling motivated. The job of a manager is to create such an environment by building trust and guiding the team with an inspiring vision. The employee feels motivated to perform at high levels because they truly enjoy what they do.
• Removing roadblocks. When working with top talent, one needs to create an environment where the employees are able to focus on the things that matter and are not distracted by bureaucracy and busy work. Though it sounds simple and seems obvious, removing roadblocks is one of the simplest ways to improve the welfare of employees.
• Help the team prioritize. Providing frameworks to help team members prioritize their work is another valuable skill that good managers possess. Recently, product developers have embraced a test-and-learn culture where moving fast is essential for progress. Overload of tasks in a fast-moving environment has made prioritization essential. Rather than help each team member with prioritization, it is most valuable for managers to provide frameworks that will enable prioritization. Generally speaking, the time it takes to execute on an initiative, the impact on the bottom line, the resources needed (and available) to accomplish a task, and the opportunity cost of doing a particular task are all important considerations in prioritization.
• Quality of analysis. There is never a point where an analysis is ever fully done. Knowing what quality of analysis is needed for a specific use case is valuable before venturing deep into an analysis. An analysis that is providing directional help in driving strategy need not be that precise, but, in contrast, an analysis conducted for the earnings call would need to be airtight. A manager should have good judgment and provide clarity to their analysts about what is needed.
• Supporting work/life balance. Managers need to avoid burning out their team members. Good managers need to keep tabs on their team member’s workloads and their overall life needs. Sometimes managers need to consider moving an employee to a different team or role to avoid the company losing a talented person.
• Finding opportunities for learning and development. Employees stay at companies when they feel that they are happy, learning, and progressing in their careers. Proactively identifying the right learning and development opportunities for each team member will help long-term retention.

Drive Excellence

Scaling should not come at the cost of performance quality. Throughout the process of scaling and building teams, there must be a culture of high performance for an organization to achieve excellence. Ideally, every employee should strive for excellence in everything that they do. If everyone has a clear understanding of what excellence looks like and holds themselves and each other to a high bar, strong company performance will follow. It is up to managers to instill this culture and set the bar.

• Setting a high bar. In order to drive a culture of excellence, managers need to set a high bar. A high bar cultivates a mindset that one is always 1% done in absolute terms. But in relative terms, more has been accomplished now versus a year ago, which provides a feeling of growth and progress.
• Don’t rest on your laurels. Building a culture of consistent improvement is critical for a high-performing organization that focuses on excellence. Certain types of individuals (A+ players) that have a growth mindset love the challenge that comes with constantly striving for excellence.
• Turning good analysts into great analysts. Somewhat counter-intuitively, managers should not spend time equally with all of their reports. Instead, they should invest time disproportionately in their rock stars. These top employees have the potential to become great analysts and future team/company leaders. Driving excellence and pushing their limits by transforming them from good analysts to great analysts should be a good manager’s focus.
• Implementing the right processes for excellence. Processes are valuable to raise excellence and scale a team. There are multiple processes that a manager could incorporate on excellence related to coaching, mentorship, code review, analytic review, etc. For analytic review, the collective intelligence of multiple people is much greater than any one person. As product analytics is still a nascent field, having a process on what great analysis looks like and having multiple people review the analysis with the focus on setting a high bar and improving the quality is valuable. Specifically, having an early meeting at the brainstorming stage of any analysis and final stages when the analysis is mostly finished are valuable points where analysis may be reviewed.
• Instilling a culture of excellence. A culture of excellence is propelled by competing with oneself rather than with others. Competition between people can erode trust and create a toxic environment over time. A culture of feedback, self-awareness, and competing with oneself creates an environment where excellence and setting a high bar are valued.
• Matching people with the right roles. People perform at the highest levels when they are good at what they do and love what they do. That’s why getting the right people in the right roles matters. Instead of a one-size-fits all approach, managers should identify what employees are passionate about, what they are good at, and what their goals are, and match them with the roles that best suit those skills, abilities, and interests.
• Quality of analysis. There is never a point where an analysis is ever fully done. Knowing what quality of analysis is needed for a specific use case is valuable before venturing deep into an analysis. An analysis that is providing directional help in driving strategy need not be that precise, but, in contrast, an analysis conducted for the earnings call would need to be airtight. A manager should have good judgment and provide clarity to their analysts about what is needed.

Product Leadership

Data team managers wear two hats, a functional hat and a product hat. Wearing the functional hat involves thinking about how to make analytics the best it can be. Wearing the product hat involves addressing how to provide the best direction for a product. As an analytics manager becomes more senior, they start to resemble a product manager. They also participate in scaling the analytics organization overall. The level of product and organizational leadership is what distinguishes senior leaders from junior ones. With respect to product development, the role of a data team manager is to help drive product strategy, set roadmaps, and grow from a functional leader into a product leader.

• Drive product strategy. A manager should be involved in shaping product strategy and vision, promoting ideas, and building consensus for the team. A manager who is able to drive strategy using data will help elevate the function itself. Senior managers are in a unique position to oversee multiple areas, so it is their responsibility to understand how those areas are connected and suggest strategic areas for improvement. For example, using data that shows engagement drives growth and recommending increased content production as a result is an example of a data-informed strategy.
• Own and drive cross-functional efforts. As a manager becomes more senior, they need to own and drive initiatives that are cross-functional. This helps them transition from being purely functional leaders to being product leaders.
• Set roadmaps for product. One of the most leveraged activities of analytics is to drive roadmaps for products. Data scientists are generally good at performing interesting analyses. It requires a different muscle to transform these interesting insights into actionable ones that lead to a roadmap for development. This requires data team managers to invest in individual coaching and building the right processes to help data scientists evolve.

Scaling Teams and Organizational Leadership

Scaling teams is more an art than a science. Fostering a happy team that focuses on excellence and execution, that is prioritizing in a way that maximizes impact is the only way to sustain strong teams over the long term. Here are a few things to consider as managers try to scale their teams:

• Optimal organizational structure. Organizational structure plays a strong role in scaling a team. Having too many managers with few individual contributors or vice versa will fail in the long term. Hiring too slow or too fast also is problematic. Ensuring that the ratio of data scientists to product managers as well as the ratio of data scientists to data engineers is optimized requires careful thinking and foresight. Being deliberate on the organization structure for both current and future needs is invaluable.
• Develop visionaries and strategists. For a company to scale, it needs the presence of visionaries, strategists, and operators. Visionaries can think big and create a compelling vision for the company. They can also clearly articulate how the world is trending, how products are changing, how people are behaving, and how the company’s product and mission tie to the vision. Strategists are able to continuously evolve the company’s direction. The role of the manager is to identify these people, develop them, and maximize their impact.
• Focus on execution. Companies generally don’t fail because of a lack of vision but rather a lack of execution. For a company to be truly successful, it needs excellent operators to follow through on the goals and roadmaps that support the vision. These operators execute and build a great work culture and ensure that products get built at the highest quality in the timeliest manner with the minimal number of employees.
• Identify and develop future leaders. At least part if not all of the next generation of leaders needs to come from within the company. Identifying these employees at an early stage and providing the support and mentorship they need to become future leaders is the role of a manager.
• Ensure that at least a quarter of a data team is made up of rock stars that can anchor the team’s work. Hiring A+ people and retaining them should be the focus of any manager. The future of the team depends on the quality composition of those rock stars. Their attitude is infectious, they set a high bar, and make everyone around them improve. They also help managers themselves elevate and transform into product leaders.
• Develop a playbook. The litmus test for a great manager is how well they are able to use process to scale multiple self-sustaining teams successfully. It is never a good idea to build multiple teams when one is not able to get one team to be great. Helping identify entirely new teams, projects, or focus areas and building the teams necessary to execute on those ideas requires the development of a framework that can be used in many situations like a playbook.
• Analytics culture. Data analytics is a nascent field, so data team managers need to consistently make culture or organizational changes that have an impact across the entire analytics function within the company.

TAKEAWAYS

• The role of a manager is two-fold: building a world class team and driving impact.
• The manager needs to focus on keeping employees happy, driving excellence, and scaling the organization, as well as on providing thought leadership for products.

This work is a product of Sequoia Capital’s Data Science team. See the full data science series here. Please email data-science@sequoiacap.com with questions, comments and other feedback.