How to scale analytics at your organization: Insights from data leaders

Data is as central as any other pillar of your business. Scaling data analytics to an adaptable, enterprise-wide data architecture demands more than raw technical skill. Organizations need an integrated approach to analytics to manage data complexity at scale.

Our guide, How to do analytics at scale: 10 tips from data leaders, is a curated collection of insights from seasoned data leaders, each sharing practical strategies and guiding principles that have proven essential to scaling analytics at their organizations.

From fostering transparency with non-technical stakeholders to structuring teams with DevOps principles, these tips reveal how organizations can avoid common pitfalls and instead empower teams across the business to work with data.

Here’s a sneak peek.

Tip one: Start with business impact, and design your team accordingly

In the past, data teams were often isolated from the business side, focused solely on collecting data while leaving its interpretation to others. Today, though, integrating data teams into specific business domains transforms their role: they become close, strategic partners within the business. Embedding data experts in each domain allows them to align directly with the business context, ensuring that data is both relevant and actionable.

When data teams are embedded, they know exactly which data to collect, how to transform it effectively, and how best to present insights that directly solve business needs. This proximity to business users unlocks a more substantial, measurable impact on business outcomes, driving data’s value far beyond simple collection.

–Raman Singh, Engineer Manager, Analytics, Symend

Tip two: You’re more versatile than you think

When it comes to scaling data organizations, data leaders should know that analytics engineers are so versatile. They have the business context of analysts, but they’ve also picked up these more technical engineering skills. They can be a bridge between data analytics and software engineering, but it goes further than that. As you scale, lean on your analytics engineers and their skillset to flex into infrastructure problems that might traditionally call for a DevOps engineer. Analytics engineers have the skills to solve their own infrastructure problems. Embrace moving your data team up the stack.

–Katie Claiborne, Founding Analytics Engineer, Duet

Tip three: Lean on DevOps principles

The first thing to stop doing is thinking about technology. It’s not about technology. It’s about people, processes and then technology.

The first thing to address is the problem of how disconnected people can start working in a more connected way. We had multiple teams using different technologies, but more importantly, the ways of working were different. We had teams working in waterfall, we had teams working in agile, working in one-week sprints, other teams working in three-week sprints. How do we get these people all on board into the same ways of working?

DevOps—or its spinoff DataOps—is a must have nowadays. It’s easy to build new solutions. What’s hard is to maintain and scale those solutions in the long run. If you don’t have a solid DevOps or DataOps process in place, you’re not going to go far. You need to change the culture to embrace DevOps.

And then obviously technology comes into the picture. We wanted something that is code based because for us CI/CD was non-negotiable. Why? Because it embeds reliability into the release process and ultimately enables scaling. On top of that, even though our scattered data teams were using different technological stacks, all of them had something in common—everyone knew SQL. But again, it’s not only about technology.

It’s really about getting people and process in place and then thinking about the technology. It takes a lot of convincing and many after-hour meetings (especially if your teams are spread across four different time-zones). But eventually you will start to gain momentum and start scaling. Once we had a solid foundation in place, it became easier and easier for us to go to a new country and say, “These are our ways of working and here’s why you should accommodate these in your lifecycle.” There was a clear tangible benefit.

Now we have all the workstreams deploying their production- grade pipelines on top of our data platform. The tech leads of each of these teams come together in the same sprint planning sessions, in the very same stand-up sessions, and in the very same sprint retrospective sessions. We guarantee that we are aligned on the roadmap and execution, and that we don’t reinvent the wheel.

Even though we’re a large team made up of many smaller project-specific workstreams, we deliver together as one unified team every two weeks—not every six months—and this in itself was a big mindset change.

–João Antunes, Lead Engineer, Roche

Download the guide for the rest of the tips. As you dive into this guide, consider these elements of a roadmap for the processes, tools, and mindsets that can transform your organization’s approach to analytics at scale.