How To Build & Manage a High Performing Data Analytics Team

CEO, Portable

Have you been hired to build a data team? Are you looking for a step-by-step approach to your first 2-4 weeks on the job?

Ben Rogojan (aka The Seattle Data Guy) and I have collaborated to pull together a step-by-step guide. You can check out Ben's YouTube summary here!

Step 0: Don’t fail

We want you to succeed.

So does your team, your manager, your CEO, and your company’s board of directors.

Unfortunately, it’s easy to fail as a Head of Data.


  1. Solving problems that don’t matter
  2. Spending too much money
  3. Moving too slow

From our own first-hand experience, we’ve outlined 10 simple steps to go from zero to success as a Head of Data.

Let’s get started.

Step 1: Figure out why you were hired

There are only 3 mandates for a data team - analytics, automation, and product development
There are only 3 mandates for a data team - analytics, automation, and product development

To figure out the answer, take the projects you’ve been tasked with, the job description you were provided during the hiring process, and the success criteria that your manager has laid out for you, and categorize them into the three buckets above.

Then pick one.

Comparing the strategic value of a dashboard, to the money saved by automating a workflow, to the revenue generated by a data product is comparing apples to oranges.

Always pick one priority when standing up a data team.

Is the answer not clear? Maybe you were hired for reason #4. To figure out which bucket to start with.

Make sure you’re aligned with your boss and leadership team on the single most important priority for your team.

In most cases, the answer is analytics.

Step 2: Start with analytics

In most cases, the fastest way for a data team to add value is through analytics - i.e. building dashboards to inform strategic decision making.

If you’ve been hired to automate workflows, or build data products, you should still understand the power of analytics, and how to approach the problem.

Where should you start?

  1. The technology doesn’t matter
  2. The team you have doesn’t matter
  3. SQL vs. R vs. machine learning doesn’t matter
  4. Whether you use a data warehouse or a data mesh doesn’t matter

First and foremost you need a curated list of the most impactful business problems to solve with data.

Step 3: Find the highest value business problems

The easiest way to identify the highest value projects is by meeting with your leadership team.

Before each meeting, spend time researching the best way to run each business unit with data. What metrics can you use to run an HR team? How do you measure conversation rates throughout the sales and marketing funnel? Which KPIs does your CEO track today? Etc.

Don’t feel comfortable talking to your leadership team yet? Talk to all of the people that report to them to generate ideas.

Once you have a sense of the data you could use to run each business unit, get time with your CEO, CRO, CMO, CHRO, etc.

During each meeting ask the following questions.

  1. What Key Performance Indicators (KPIs) do you use to run your business?
  2. Where do you find them?
  3. What metrics do you wish you had at your fingertips every morning?
  4. What is the action you take based on each?
  5. How do you measure the impact on the business? Is it critical?

The last two points are critical.

The goal of analytics is not to present data for the sake of data. The goal is to inform actions that have a material impact on the business.

Step 4: Summarize the opportunities

Here’s an example.

In talking with your CMO, you might identify the following:

  1. The number of blog posts posted each week is a KPI.
  2. It is manually tracked by the Head of Growth in a spreadsheet.
  3. The CMO not only wants to know the number of blog posts by week, but also how many people are viewing each blog post.
  4. Both metrics should be increasing each week - the number of blog posts, and the total engagement - if they are not, the CMO will put time on the calendar with the Head of Growth to discuss ideas for improvement.
  5. Your company uses product led growth to drive sales, so 100% of your paying customers have come in as a direct result of content engagement. Having these insights automated will allow the CMO to make sure the company increases lead generation from content, and therefore revenue, by 40% in the next year. It’s very important.

Once you have a list of possible analytics projects. Rank them by perceived value.

Then, go back to your leadership team to make sure they’re aligned on how valuable each project is and see if they agree with the prioritized list you came up with. This allows you to make sure you’re focusing time only on the most impactful projects.

Step 5: Identify the necessary data

For each initiative, you now need to understand what types of data you will need access to..

If you’re building blog post insights, you’ll need: metadata on blog posts, at what time they are posted, and engagement metrics over time.

For each initiative on the list, write a brief summary of the data you would need to solve the problem.

Step 6: Track down the data

Next, you need to know what data is accessible.

To identify what data is accessible across your organization, you should do two things.

First, set up meetings with mid-level managers across the organization to identify which systems their team uses. They typically fall into three categories

  1. Databases - Mostly in use by engineering organizations or product teams
  2. Business applications - Used across the organization
  3. Event data sources - Used by marketing and product teams to track usage on websites and mobile applications

Second, talk to IT - They can use a single sign on tool to pull a comprehensive list of tools. They also manage procurement of new vendor solutions, so they are tapped into which tools are being used at scale across the organization

Once you have the list of tools, do some research into what types of data exists in each tool. Is it customer data? HR data? Sales training data?

Go back into your list of high value analytics projects, and identify which data sources could solve each problem.

Step 7: Pick the highest ROI problem

Start at the top of the list, and figure out which ones are feasible.

Are there high value projects that you need to put on hold because your company doesn’t collect the data? Skip them for now.

Are there any easy wins? Where the value is high, and the data exists? For example, your CEO wants to know how many monthly active users your product has, but no one has built a dashboard on your product data? Easy win!

Find the highest value project you can accomplish, write it on a Post-it note, and tape it to your forehead.

It’s the only thing that matters.

Step 8: Build the minimum viable tech stack

Don’t waste your time on infrastructure. If the goal is to create an automated dashboard to track blog post delivery and engagement, you only need three things.

  1. A visualization tool that your CMO can leverage
  2. An ELT tool to pull data from your blogging and engagement platforms
  3. A warehouse to store the data

See if any of these already exist. If so, use those to start.

If not, ask for a budget.

Worried about the budget not getting approved? Don’t be.

You are working on one of the highest value analytics projects at the company, your leadership team agrees that it is a high priority, and you have justification from the CMO that it is critical to the business.

Infrastructure for the sake of infrastructure deserves no budget. Infrastructure necessary to solve the most important business problems is an easy investment.

Step 9: Solve the problem

Buy the tools you need, connect to the data sources, load the data into a warehouse, and build the dashboard.

The first time you solve an analytics problem, the technology, people, and processes only need to work for that problem.

For problems 2, 3, 4, 5, etc. you will need to think through adding capabilities, scaling processes, and expanding the team.

Do so incrementally. Use an agile methodology. Don’t buy things you don’t need. And don’t get stuck on infrastructure for the sake of infrastructure.

Step 10: Repeat

You should think about the items outlined above as a framework:

Every year, you should revisit

  1. What is your job description?
  2. Is analytics still more important than automation and building data products?
  3. Is your team and mandate big enough to expand scope?

Every quarter, you should revisit

  1. What are the highest value data projects?
  2. What data do you need to solve the problems?
  3. Does the data exist in your organization to solve these problems?

Every month, you should revisit

  1. Which is the highest ROI problem to solve with data?
  2. What is the minimum viable tech stack to solve the problems you have identified to-date?
  3. How do you solve these problems?

Every day, you should stay away from

  1. Solving problems that don’t matter
  2. Spending too much money
  3. Moving too slow


Building a successful data team is hard work.

You set goals, wrangle stakeholders, identify projects, prioritize work, and deliver value.

By sharing our perspectives, we hope to accelerate your journey of building the next great data team.

Best of luck.