Switching Away from Google Analytics to Mature Our Analytics Culture at Unity

Insights/Action/Outcome: With as many as 35 properties attached to a single project, Unity found they could collect richer data with Amplitude than with Google Analytics in a more flexible way. That allowed Unity stakeholders to perform more sophisticated analysis and understand how their work ties to other stakeholders' objectives, leading to better collaboration and insights.

Sunsetting technology can be a blessing in disguise since it forces you to take a fresh look at your tools. At Unity, we took a routine rollover to the latest version of existing technology and turned it into a new direction for the company.

I’m the Senior Data Engineer, Measurement Strategy & Analytics, at Unity Technologies. I’ve been at Unity for two and a half years. My entire career has been based mainly on implementing analytics tools for enterprise-sized environments. These are complex companies that operate across several countries, languages, currencies, and e-commerce platforms.

I help stakeholders and teams define what’s important to measure and then provide the right tools to collect, analyze, and display that data in dashboards and reports. Improving measurement culture in management is a big part of this role. Trying to persuade people to do things a certain way, even if it will help them, can be complicated. And at a company of Unity’s size, it can be tough to convince teams of busy product managers and engineers that analytics should be part of their processes.

Unity used to rely on Google’s Universal Analytics. It made sense since it was a standard, so to speak, and it integrated with other Google Advertising platforms. But there was something I didn’t like about Universal Analytics: Even though GUA pageviews and events (category, action, label) were events in nature, you had to work within Google’s predefined events, and they were not flexible enough. That limited framework didn’t give us the flexibility the business needed. Then Google announced they were sunsetting Universal Analytics and replacing it with Google Analytics 4. I’m not a great fan of moving to the next thing a vendor throws our way just because it’s there. It was the perfect moment to look at alternative options to find the right solution for Unity.

Don’t just rethink the tool: redefine the implementation

Unity needed an enterprise tool, which usually comes with an enterprise price tag. There is a second tier of analytics products that are a little bit cheaper but don’t have the flexibility you may need at the enterprise level.

Amplitude was different: It had all the sophisticated capabilities an enterprise needs at a reasonable price. In contrast to Google Analytics’ constraints and shortfalls, Amplitude is 100% flexible. It would allow us to not only define events the way we want without the erratic and oftentimes unfortunate product roadmap GA4 has.

At the end of the day, everything is analytics for me: I don’t distinguish much between product and marketing analytics, though there are certain features a tool has to nail if it wants to be competitive in both areas.

From the marketing team’s perspective, the typical product analytics tool doesn’t have robust enough attribution and e-commerce tracking to replace Google Analytics, for example. Some of those features need a bit of polish, but the Amplitude team listened when I provided feedback. And they continue to listen. The feeling that somebody’s supporting you? That’s something I value because there’s always going to be some bumps in the road. How your partner approaches those issues and resolves them is what matters.

I was leaning towards Amplitude when I learned that some Unity teams were already using it for product analytics. This made the Procurement process simpler since we would only need to extend Unity’s existing contract to include more events. In December 2022, that’s what we did.

Our biggest migration challenge was time. We had a deadline of six months, when Universal Analytics would stop processing new data. On the other hand, we had done a lot of previous work to get things right, ensuring we were scalable and without dependencies. When Google forced these migrations, many people were upset at the idea of having to re-implement everything from scratch. Not me. Rethinking your implementation strategy is an opportunity to seize—not something to fear. It’s the chance to ensure you’re meeting your company’s present and future needs.

When I talk about the strategy of proper implementation, people don’t always get it. If you just throw down the code and start tracking, you are going to suffer when it comes time to make the changes that come with the growth of your product or the tools you use to measure performance. Only planning with scalability in mind will ensure long-term success. That’s what we did, and by early March we were ready to switch off Google Analytics and turn on Amplitude.

The impact is how we’ve matured in our analytics culture

Every team at Unity now uses Amplitude. One of the things I like about Amplitude is it’s very simple for a person to create a dashboard with only the data points that they need. On the opposite end, advanced users—those who deal exclusively with databases, and never go near a user interface—can spend their days in their database of choice, like BigQuery, fueled with automated imports from Amplitude, running complex queries and using data pipelines to merge data from many sources. Users of varying technical proficiencies all have better visibility into the data, in a way that makes sense for their use case. That’s the key to data democratization and a major contributor to Amplitude’s effectiveness within our organization.

Not only that, Amplitude has furthered Unity’s experimentation culture. Using Amplitude Experiment alongside Analytics is seamless, which means our teams don’t have too many tools to deal with. Unity teams now use Analytics for tracking and Experiment for feature flags and feature rollouts to fine-tuning products.

My goal at every company is to mature the company’s measuring strategy, aiming for more advanced applications of data. A few years ago, everybody was talking about personalization. It’s super cool, sexy stuff, but you have to be more mature in your technical backend if you ever want to get there. Personalization requires more than an interface that can provide personalized options to a user—your backend needs to be involved to provide key information.

In our first year after implementation, Unity teams have taken a deep dive into Analytics and Experiment for their daily work. The marketing teams create user cohorts to improve the effectiveness of their campaigns, for example. But we’ve only scratched the surface.

One of the biggest benefits that drew me to Amplitude is its ability to define events using multiple properties. For example, a call-to-action button will mean something different to a marketing person than it will to our UX team. Defining a single event in multiple ways or with many different properties means I go into meetings asking teams what they want to measure, and I have the ability to actually deliver.

On an organizational level, being more mindful in defining events provides a structured and controlled approach to data collection and analysis. Not everyone sees the value of that. Give some people the option, and they will collect all the garbage in the world. But from my perspective, I love how Amplitude’s Data Governance ensures data quality, reducing the need for post-collection cleaning, and ultimately makes analysis that much easier for the end user.

What my colleagues now see in the data

My role focuses on giving teams the data they seek and the ability to transform that data in ways they find meaningful. Selecting the right tool and implementing it properly as part of a larger strategy is like delivering my organization to a safe harbour. I don’t always get to see the fruits of my data labour, because I have to focus on the delivery as the complex, ever-changing scenario I deal with keeps me busy enough.

What I do hear from teams is they realize we are collecting richer data than before. With lots of properties attached to a single event, teams perform more sophisticated analysis than they could under the constraints of Google Analytics, and that allows them to pinpoint areas for improvement.

Richer data also improves collaboration between teams. When you can see that an event becomes meaningful to someone in another department, you think, “We should have a conversation about this.” Suddenly teams stop operating in silos. Easy-to-use tools also grow a company’s measurement culture faster. Amplitude’s user-friendly interface and adaptability allow teams to experiment and glean valuable insights without becoming tangled in technical intricacies.

My advice to anyone considering moving to the latest version of their existing analytics tool is to see if it’s the best option for your company. Take the time to explore the alternatives and what they offer both in functionality and as a partnership. There is always a cost to migrate, but sometimes you have to migrate anyway, so your obligation as a professional is to make the best of the opportunity. No solution is going to be perfect for everybody, especially considering the complexity at the enterprise level. Your job is to have the curiosity and adaptability to find the best solution for your organization’s unique challenges now and into the future.