Amplitude switches to DX for engineering insights
Amplitude, a leading digital analytics platform, has long sought to understand how engineers work as part of their efforts to improve developer productivity. Early on, the company turned to an external vendor in an attempt to understand where engineers were spending their time. “That Vendor was implemented long before I joined,” explains Ryan White, Senior Director of Infrastructure, Platform, Security, FinOps, and Developer Experience. “At the time, it solved the problem Amplitude had, which was getting basic productivity data out of GitHub and Jira, but when I took a fresh look at what the other vendor provided, I quickly realized we had outgrown their capabilities.”
The central issue White saw was a lack of customization and configurability. The other vendor forced teams into fixed dashboards and rigid interpretations of engineering data, with no access to the underlying queries. “Our biggest challenge with the other vendor was a fundamental lack of self-service and flexibility,” White explains. “All you could do was take their dashboards and the way they chose to interpret them. There was no queryability or AI chat agent, and there was no way to build custom reports or dashboards, despite a number of feature requests.”
Basic team configuration changes also required filing a support ticket with that vendor. “Every time I wanted to add a new team, I had to file a support ticket, and I could not make changes to team filters without a support ticket. It was toilsome.” For a leader used to shipping fast and iterating quickly, this was a major blocker. Beyond the self-service limitations, Amplitude needed integrations that the other vendor wouldn’t provide. White wanted to pull in data from additional sources like PagerDuty to get a more comprehensive view of engineering quality and on-call burden, but these integrations weren’t available.
When Amplitude began evaluating alternatives, White conducted a head-to-head comparison between the other vendor and DX. “I’ll be honest, it wasn’t close,” he says. Adding data to DX was simple. “When we trialed DX, we were easily able to port in what we had, so that it was kind of apples to apples,” White explains. "We weren’t suddenly creating a whole new set of measurements.”
Beyond replicating what the other vendor provided, DX offered the flexibility and customization that Amplitude had been missing. DX gave White and his team the ability to query data directly, build custom dashboards, and integrate additional data sources. “DX was able to pull from additional data sources that the other vendor didn’t have,” White explains. “We connected PagerDuty, for example, and started to include PagerDuty analytics in our quality metrics. We also added webhooks when we deploy to production on Kubernetes. Anytime there’s a production deployment, it calls DX, so DX knows we deployed this workload, and we can correlate with DORA and other metrics and reports.”
The ability to create custom reports has also proven to be particularly valuable. “I created a custom security dashboard that shows how we’re responding to and triaging security discoveries from our array of D&R tools versus our SLA, because we have a lot of enterprise customers that demand strict SLAs on security,” White shares. “I’ve also created a custom dashboard that tracks all of our incidents by team, pillar, pod, and service, so people can really drill into what the contributors and trends are.”
Today, DX is Amplitude’s central hub for engineering intelligence. White describes it simply: “We use DX to collect every ounce of metadata from all the tools our engineers use, whether it’s GitHub, Jira, PagerDuty, Cursor, Claude, or Copilot to analyze productivity, time investment, and AI Impact. It allows us to tie all this data into one place and run queries on it to understand where we can improve.”
More recently, Amplitude is also using DX to guide its AI initiatives by correlating AI usage to improvements in DX metrics in order to understand how AI tooling can accelerate work at Amplitude. “We needed something that could actually show us the impact,” says White. With DX, the team can correlate how AI tooling influences cycle times, code quality, and developer sentiment, compare those signals across teams, and track changes over time. Instead of relying on assumptions, Amplitude now has data on how AI is shaping engineering performance and where future investments could yield the greatest benefit.
Finally, another capability Amplitude has found powerful is DX’s support for custom data ingestion and analysis. Using DX’s Postgres tables and connection string, Amplitude can pull its own datasets directly into DX and query them alongside DX’s metrics and data. For example, White pulls in meeting event data from Amplitude’s Engineer IC calendars, maps it to teams and roles using their org chart, and calculates signals like focus time, context switching, and weighted calendar efficiency. By sending that data into DX, White can now explore correlations between calendar efficiency and engineering productivity in a single, unified view.
While the self-service and flexibility were the primary drivers for choosing DX, White found an added bonus after making the switch: the speed at which DX turns around feature requests. “At first, I thought the fast turnaround was just because we were in a trial and DX wanted to win our business,” White says. “I was happy to see that DX has continued to turn things around pretty quickly.” That experience stood in stark contrast to the other vendor. “There were four or five feature requests I filed with them, and I waited nine months while checking in regularly, but there was just zero activity on them,” says White.
Without DX, the work of understanding engineering at Amplitude would be significantly more labor-intensive. “Life without DX would be far more toilsome,” says White. “DX has made it very easy for me to ask a question about engineering at Amplitude, and then go find the answer.” With Amplitude’s engineering organization continuing to scale, DX will play an essential role in shaping and validating future investment decisions.
