The Analytics Skill Matrix is a tool we use in Data teams throughout Automattic. It developed from our interest in creating clearer job postings for hiring and detecting the company’s technical and non-technical growth areas. As such, it embodies our Automattic Creed’s reminder to “never stop learning.”
We have used the Skill Matrix for many purposes:
- Develop job postings;
- Evaluate the strengths of data professionals;
- Evaluate people’s potential areas of improvement;
- Assign projects to people based on their strengths;
- Use the information for hiring to complement our knowledge gaps;
- Decide and pick learning programs for the team;
- Pair people for learning;
- Assign people to projects based on their complementary skills.
The tool can be handy for data-related roles such as Data Scientist, Analytics Engineer, Data Engineer, Data Analyst, Analytics Engineer, and more. The Matrix can also help you, as an individual Automattician, discover your strengths and areas where you may wish to improve. This increased awareness is valuable when planning new learning opportunities and applying for jobs, as you can focus on developing the most relevant skills for the type of role you want.
For the current project, we designed a spreadsheet by making an inventory of two dimensions, our needs and competencies—that is, the tasks we need to perform to fulfill our mission and the skills plus tools we need to complete those tasks—which, combined, let us perform our analytical duties.
Let’s take the mission of Tumblr’s Data Science Analytics team, “Supporting the business by providing accurate and timely metrics and insights to drive data-driven decisions.” To achieve it, we must first assess the dimension of the need as a team. Some of our needs are:
- Designing and implementing relevant metrics and KPIs to monitor the business;
- Reporting and presenting the state of the products across the business;
- Aiding in making data-driven decisions by performing data analysis;
- Experimenting with products and fostering their growth through data.
With these needs in mind, we select tools and develop competencies to fulfill them: Some tools we habitually use are SQL, Python, Looker, Apache Airflow, PySpark, Hue, Hive, Jira, scala, PHP, Kafka, Java, etc.
We created a list of more than seventy competencies by considering our needs and the relevant tools. Then we categorized the skill expertise into the four levels we use internally: newbie, basic, comfortable, and expert. Finally, we combined the list of competencies and the level of expertise to design the skill matrix.
Among the list of competencies created, we expected the ability to…
- Perform insightful deep dives (Exploratory Data Analysis);
- Develop Dashboards;
- Define KPIs and metrics;
- Interpret the results of A/B tests;
- Design experiments;
- Define requirements from vague statements;
- Provide reliable estimates for assigned tasks;
- Take on tasks and projects from start to completion;
- Coordinate tasks between multiple stakeholders;
- Establish rapport and positive professional relationships with team members;
- Build data pipelines;
- Write and present the results of experiments and analysis clearly.
It also helped us to think about the following categories when creating the list of competencies:
- Technical competencies;
- Domain knowledge;
- Project management and execution;
- Teamwork and communication;
- Leadership skills.
Once you have your list of competencies, you can add the horizontal expertise dimension to them to produce the final skill matrix. On the vertical axis, you would have the competencies, and on the horizontal axis, the level of expertise in your desired scale. You can see below a small example of some of the skills we measure for our purposes:
|Perform insightful deep dives (Exploratory Data Analysis)||x|
|Design and develop dashboards and compelling data visualizations||x|
|Develop efficient data pipelines||x|
|Design and analyze A/B testing experiments||x|
|Communicate clearly technical topics to non-technical audiences||x|
One beneficial way anyone can use a skill matrix is to perform a self-assessment of their current skills. I did this when I wanted to level up my technical skills as a Data Scientist upon joining the team. I reviewed the skills I had and my expertise in each one, and compared them to the needs of our team. This way, I discovered that, even though I was already an expert in SQL, I didn’t fully understand our data pipeline system since it was based on Apache Airflow, a tool I didn’t use often.
I followed up by requesting feedback for resources to learn Apache Airflow for data colleagues in Automattic. Then I bought some books and invested many hours studying, designing, and developing data pipelines. The Skill Matrix, and the insight it provided me, enabled me to make valuable contributions to our team’s data pipeline and increased my understanding of our data systems. I have used this method many more times to “never stop learning” and focus on impactful development areas. It has helped every time.
Elaborating accurate job descriptions
It is not uncommon to see unicorn-like skill demands in job postings, where it seems they are requesting multiple roles for one person. These job postings usually arise due to a problem of inaccurate needs assessments, resulting in a delayed and complex hiring process. As candidates with only a small set of the demanded skills might be discouraged from applying.
Another beneficial way of using the Skill Matrix is to assess the team’s needs and align the skills in the job description with those needs. One would use the Skill Matrix to understand where more expertise is needed and write a job description that tailors to those particular skills, resulting in more detailed and accurate job descriptions and a more efficient hiring process.
Our Skill Matrix is a tool anyone can use for many different purposes, and one that is relatively easy to build once we consider dimensions such as competencies, tools, and needs. Through the exercise of designing one, you will gain an increased awareness of yourself, and improve your ability to plan for your objectives and prepare for data-related roles.