Managing Growth in the Analytics Organization
Building an Analytics team in a startup in Brazil has its challenges. While our ecosystem is receiving unprecedented amounts of venture capital, the tech community in Sao Paulo is a child compared to more mature hubs in the United States or Europe, where tech companies thrive for +40 years. We have a scarce senior talent market.
Growth Mindset
Since the beginning, we have adopted the growth mindset at Liv Up. We knew that to achieve our mission, our young and inexperienced team would have to grow.
At a fast-growing company in a talent scarce market, developing people is not only a desire but a need. To accelerate people’s growth, we built a framework based on these transparency principles:
- Clear expectations
- Constant feedback
This framework was strongly based (sometimes copied) on previous work from awesome people from other companies. Special thanks to the Medium Team and their Growth Framework, the Career Ladder Series from Locally Optimistic (I, II, III), Camille Fournier, and her awesome The Manager’s Path, #compensation, and #perf-management at Rands Leadership Slack. You are amazing!
The overarching structure of the process is: people build and develop skills, that enable more challenging roles, that change compensation.
Clear Expectations
Defining clear expectations is essential to lead a team. It lets people focus on what matters, defines milestones for promotions, and avoids frustration.
Expectations provide structure for feedback. If it is not clear what is expected from someone, it is hard to provide actionable feedback. With clear expectations, providing feedback is easier.
Career Ladder
The Career Ladder defines our career paths, roles, and responsibilities.
The first step was to define the Analytics mission and main responsibilities at Liv Up. The Analytics scope varies greatly from company to company and we had to get it clear.
Mission: Promote data-informed decisions throughout the organization
To achieve our mission, we have 4 main responsibilities:
- Coordinate the data capture process
- Transform data (SQL in Dataform)
- Metric design and visualization (LookML and Looker Dashboards)
- Analysis and presentation
From these responsibilities, we could define the main activities and define our Career Ladder.
We followed the standard structure of two paths: individual contributor and people manager. From level 1 to 3, there is a core path, where all analysts should develop technique and execution skills. On level 3, team members decide which path to follow.
As the responsibilities of individual contributors and people managers vary greatly, we built a RACI Matrix to communicate what is expected from people managers. For individual contributors, the main differences are the depth, complexity, and impact of their projects.
After acknowledging what the company expects from the Analytics Organization and defining what is expected from each role, we defined which skills we wanted in the team.
Skill Snowflake
When studying classical career ladders we faced a problem: Analytics is so diverse and people can impact the organization in different ways. Our Business Analyst role is a generalist. They coordinate the data capture process, model data, develop analyses, present results, and even build automatic decision systems. It is a mixture of analytics engineering, data analysis, and data science.
We wanted our organization to embrace this diversity, with each team member complementing other’s skills. Standard career ladders follow a linear path, assuming everyone will evolve in the same direction.
Then we discovered the Medium Growth Framework. A framework that enables growth in multiple directions and we decided to adapt their framework to our needs. There are 16 skill tracks, divided into groups: Technique, Execution, Community, and Support.
- Technique: Modeling & Architecture, Analysis & Presentation, Statistics & Algorithms, and Metric Design & Business Sense.
- Execution: Project Management, Communication & Influence, Craft, Initiative
- Community: Mentorship, Data Evangelism, Recruiting, Networking
- Support: Career Development, Org Design, Psychological Safety, Feedback
Each track has 4 milestones, that are illustrated with examples. Examples are essential to communicate clearly what is expected from each level.
You can take a look at our full rubric here (this document is always a work-in-progress!).
Medium developed a structured process that supports their promotions. We have a different approach, the Snowflake is a map, it helps everyone in the organization to find a direction of growth, with clear expectations for each milestone. Our promotions are sustained by another process, the performance review, that focuses on results, not skills.
We also forked the Medium Tool to visualize the Snowflake. It is a great way to interactively visualize progress.
Compensation
Compensation transparency is a hard topic. There are many factors involved: the market, company economics, cash/equity.
Although individual compensation varies with some factors, we believe that everyone should know our compensation ranges for each level. We disclosed this information internally.
This transparency helped us align expectations and initiate positive debates about compensation.
Constant Feedback
With the skills, roles, and compensation expectations set, we are ready to build feedback processes to guide growth. Feedback is what turns expectations into performance. The leader has a strong role in guiding her team to achieve their objectives.
We defined 3 processes to support growth, each one with specific cadence and output.
Career 1on1's
1on1’s are the most important meetings between a manager and her direct report. We have regular 1on1’s (usually bi-weekly) and at least once a month the 1on1 should focus on career growth. Supported by the Career Ladder and the Snowflake Rubric, the manager can provide feedback on previous deliverables, define expectations for the following weeks, and review the Snowflake rubric.
This continuous process of iteration is the foundation for growth. It is the “Get 1% better every day” idea. Having these recurrent meetings make it mandatory for managers and analysts to constantly review performance and skills, identifying opportunities, and acting fast.
Changes in Snowflake milestones are discussed on 1on1’s. Although our Snowflake process is much more decentralized than at Medium, we try to be consistent among teams and calibrate every milestone change with other leaders, so we guarantee that our bars are calibrated (avoiding misalignments in performance reviews and compensation cycles).
We follow the 5 Cs: the analyst must have demonstrated a Conscious, Comfortable, Continuous, Consistent Competency.
Performance Review
The objective of the performance review is to provide a clear evaluation of an analyst's performance (transparency) and an action plan (feedback).
This process runs bi-annually and is deeper than 1on1’s:
- The analyst prepares a self-assessment, identifying the main results, and milestone changes from the period.
- The leader prepares her evaluation, taking into consideration her view and peer-feedback (based on company-wide 360 feedback).
- The leader calibrates the feedback with her supervisor or other leaders.
- The leader sends the final performance review to the analyst.
- The leader schedules a meeting to discuss the results and action plan.
We hold this process bi-annually, and we follow a template:
- Results (main deliverables and milestones changes)
- 3 strengths
- 3 growth opportunities
- Detailed evaluation (for more examples and context)
- Result (exceeding, meeting, or below expectations)
- Action plan (suggestions on the way forward)
The process of building this document is vibrant. It requires leaders to understand other’s perceptions, hear actively, define clear objectives, and sets the structure for following 1on1’s.
Role and Compensation Review
We opted to untie performance reviews and compensation for many reasons.
With all the information from 1on1’s and Performance Reviews, our compensation review is simpler. Every leader prepares a proposal of promotions based on people's performance and motivations, company opportunities, and market compensation changes.
To advance to a new role, people must be performing above the expectations on their current level and developing the necessary skills to assume the responsibilities of the next role.
It is discussed with leaders from the Analytics Org and approved with the Compensation Team (People Org).
Notes
Hiring
As a byproduct, we could improve our hiring process. As the skill tracks are detailed with examples of each milestone, we could design a process to assess these skills and identify the seniority of candidates.
It makes communication easier between everyone involved in the process.
First, we can define what we want for a specific position. For an analyst that will work on a team that requires a lot of data wrangling, a level 3 in “Modeling & architecture” may be mandatory, while on a team with multiple stakeholders, “Communication & influence” might be more important.
In each hiring phase, interviewers identify milestone tracks from the candidate’s previous experience or tests (business cases).
Results
We have a bi-annual Analytics engagement survey (using the Likert scale) and the results were awesome (compared to the previous semester). Some questions:
- I’m pleased and happy with my work at Liv Up. Improved: 81 to 91 points.
- I receive relevant feedback. Improved: 44 to 73 points.
- The feedback that I give to leaders or peers is well received. Improved: 81 to 93 points.
Considerations
- We implemented these processes when the team had between 12 and 18 team members. It was good timing, a smaller team wouldn’t need this structure and a much larger one could be hard to implement.
- This was a year-long process of building processes. We are constantly iterating, and it couldn’t happen without the commitment from everyone from the Analytics team.
- Shortly after introducing the Snowflake, we could notice changes in behavior. The examples in skills and milestones illustrated behaviors that are expected from people, but they didn’t know (such as Networking or Psychological Safety).
- We started two bi-weekly learning meetings to share knowledge: Management Talks and Analytics Talks. In these talks, a member of the team prepares a material (usually a book chapter or articles) and we discuss the content. We had 15+ Management talks, exploring books such as High Output Management and Radical Candor, with good feedback.
- This process is running for the Analytics and Engineering teams and we are expanding to other teams.
Finals
We are very satisfied with the results so far, but there are many things to improve! We are constantly updating our examples in the Snowflake, we are always trying to improve the quality of feedback and provide better guidance.
If you liked the content, let me know your thoughts on LinkedIn!
And we are hiring! We have several open positions at the Analytics team and Liv Up.
