On Scaling Startup Analytics: Preserving Expertise and Speed

Scaling up your business, product, and analytics team is a balancing act, especially when everyone has individual OKRs and project objectives to meet while onboarding new team members.

Your organization must stay nimble, and priorities will continually shift. The faster your team grows, the greater the need for clarity and cohesion, which is a chain reaction that inevitably increases an analytics leader’s demand for more order and processes.

If you’re reading this article, chances are you’re experiencing some version of this startup growth cycle right now. The good news is that we’ve already worked through it all and can assure you there are remedies for your analytics team’s growing pains.

Our organization, Super, doubled its team size over the past year. We’re on the data and analytics team, and we needed to quickly refocus our approach from being scrappy and reactionary to doubling down on a domain team structure for better collaboration and more valuable insights.

We also aligned ownership and our specialized skill sets to different domains and business units within the organization. Finally, we cultivated a written culture where we document everything from team charters to processes, best practices, tools, and notes.

Doing so has helped us scale more effectively while assisting recruits in ramping up their productivity. Let us show you what we’ve learned and how we did it so that you can adopt these best practices for scaling your startup analytics team.

1. Double down on a domain-based team structure

Super originally started with just two business functions: Engineering and Business. As our employee headcount scaled from 30 to over 200 people, the need for different business units and team structures grew significantly.

In the beginning, our analytics managers straddled many business areas, which caused them to be reactionary when supporting requests instead of being proactive and predicting future analytics needs. Their job also required general knowledge, so they didn’t have a chance to specialize. As a result, they couldn’t build the deep business expertise and value needed to drive their potential impact on business growth.

It’s a transition many startup leaders may face, across all business areas, during a team scale-up. Knowing this challenge upfront, however, can help you better plan for it in the future.

We went through several iterations to try and create the conditions for each analytics manager to develop deeper expertise. We finally adopted a domain-based engineering and analytics team structure, which helped to scale the organization quickly. It ensured better career growth opportunities as our business expanded, as well.

We started by splitting our analytics team into business domains, each focusing on a specialized area of the Super business. Each domain is led by an analyst or group of analysts and supported by their analytics manager to foster inter-team collaboration and balance capacity.

Through this approach, our analysts benefit from the depth of expertise gained in embedded models while collaborating and learning in centralized analytics teams.

Our domain team structure mirrors the popular Team Topologies book by Matthew Skelton and Manuel Pais. Many remote-first startups have adopted the strategies in this book to improve team interactions, better track team-level dependencies, and redesign inter-team communications.

Our transition to a domain team structure has allowed for smaller, more relevant meetings at Super. As we scaled-up our analytics team, this structure became even more critical. It also saves everyone time and frustration, and there is now greater direct ownership of outcomes.

Since adopting and committing to this structure, we’ve developed more standards and best practices, including:

• How we hire new team members.
• Creating team charters to help identify who does what within the team and the organization.

More importantly, this new structure has empowered us to be more thoughtful about who we hire as we scale the business and how we keep them engaged and happy.

2. Think carefully about your analytics team’s expertise

While scaling up at Super, we’ve slowly evolved from a team of jack-of-all-trades analytics professionals to seeking unique skill sets to round out the group.

To date, we’ve added expertise in experimentation and data science, as well as attribution and Google Analytics 4.

These changes have enabled us to deliver more specialized insights while sharing information and teaching others.

Breaking down team siloes

While scaling our analytics team, there were some challenges with people feeling siloed and needing clarification about where they fit on the team as we added new hires. We soon realized we needed to pivot and respond to issues as they came up.

Even now, we face a challenge of social cohesion. We’ve put more effort into improving and being mindful of it. Overall, it remains a work in progress, and we’re focused on restoring the connection between our engineering and analytics teams that came more naturally when our teams were smaller.

To resolve the issue, we’ve held some informal in-person gatherings to be more social. We also host a monthly video call that allows team members to share what they are working on and their expertise and ideas. We’ll keep building bridges between teams.

Additionally, we’ve learned over the past year that the onboarding process could have been more relaxed for both new hires and existing employees. Our solution was to minimize the cognitive load by creating a central repository for all the skills and information a new hire must learn, which we’ll take a look at next.

3. Improve documentation and job satisfaction to balance delivery

Creating a streamlined documentation process was the cornerstone of our central repository strategy to improve employee job satisfaction and scale productivity.

Fostering a written culture

In February of this year, our leadership team set a corporate goal to shift to a written business culture where every employee must author at least one document per quarter.

This top-down initiative flowed into our individual and team OKRs, which required every team member to create and share their project’s information through written documentation.

Our documentation creation and sharing process

We first invested time and resources into the development of team charters. It helped us define who owns what within each business domain and across the organization. It also required key stakeholders and project owners to be aligned from the start, and every meeting must have a written agenda and designated note taker. Matt Culver, our Chief of Engineering Staff, explains the process beautifully in his blog post( which was part of his written culture OKRs).

Next, we created educational documents specific to each area of the business. At first, it took some coaching and prompting to get people to add to the central repository. One person is assigned as the leader in each business area. Their job is to prioritize what documentation needs to be created and when. Then, they give someone on their team the task of writing it.

As part of our definition of “done,” the entire company has been aggressive in scheduling time to create documentation and tracking projects via Jira, our central repository tool. Every data set is tagged to an owner in the system. That way, employees know who to contact on our team for more insights. It is now a searchable, self-serve solution where everything is accessible through one location.

We also developed a process to identify when content gets stale and how to update it. Each Domain Lead needs a constant feedback loop about where a new hire gets stuck with content during the onboarding process. The Lead must then figure out what’s missing and ask someone on the team to update or fix the documentation, to improve the learning experience for future new hires.

These steps require a significant upfront and consistent investment in data documentation, documented processes, and standardizing queries in git.

Improving employee job satisfaction and project delivery

There’s been a noticeable shift in our business since we started the new documentation process, and employees are very much in favor of it now.

Super’s new written culture has enabled us to provide a more effortless onboarding experience for new analysts. The documentation process has saved our existing team members a lot of time and stress during the onboarding process, as well.

We’ve come to appreciate that one upfront investment in our time to create a document is the equivalent of one new hire conversation. That document will then save each of us valuable time in the future because there’s no need to have the same conversation with every new hire. Or, at least, it’ll be a much shorter conversation where we might only have to clarify a few key points.

Likewise, when you need to do something you’ve never done before (e.g., building a connector), you only need to spend 30 minutes reading a document on how to do it rather than asking someone else on the team.

Colleagues are always happy to help, even when they must spend time away from their own work. As a team scales, making knowledge accessible via writing will prevent meeting time and confusion from scaling accordingly.

Finally, our analytics team members are more excited about their work than before we embarked on these new processes. Employee communications are now asynchronous and scalable, too.

Retention and speed are the hallmarks of success

It can be challenging to tie hard numbers to the return on investment of adopting many of these analytics team scale-up practices. That’s perhaps why many tech companies don’t undergo these significant transformations.

On the Super analytics team, we believe that retention and speed of onboarding (e.g., time to first insight) are our hallmarks of success.

The size of our analytics team has effectively doubled since the start of the year, with no turnover in the last twelve months. It is a strong indicator that the changes we’ve made to our team structures and hiring strategies significantly impact employee job satisfaction.

We’ve aligned our team’s specialized skill sets to different business units or domains, making it easier to work directly with those analysts. These strategies have helped all teams move faster towards highly focused outcomes at Super.

Overall, our analytics team feels more productive, and new hires find that they can add value earlier from day one than they would without our domain structure and documentation. That’s because it reduces the time to transfer knowledge to new hires and between teams.

New team members have told us that structured OKRs specifically detailing what they’re responsible for regarding documentation are helpful. They also like to have a written outline of what’s expected of them within the first 30, 60, and 90 days after joining the company.

Finally, our written culture has become more organic and self-reinforcing, as employees want to document everything because they’ve seen and experienced its many benefits.

Have additional questions about scaling analytics teams within a tech startup? Please leave a comment below.

If you want to join the Super team — check out our careers page.