What I like most about GoodData.AI is the tools that govern the semantic layer. Those features have delivered the most measurable impact for how we operate at scale. The flexible dashboards and robust data integration allow for seamless visualization and analysis of large datasets from multiple sources. Beyond that, defining our key metrics once in the semantic layer and having every client dashboard inherit those definitions automatically has been the single biggest time saver since we went live. Before GoodData.AI, we had client-facing reports recalculating metrics independently, and discrepancies between what a client saw and what our internal team saw came up roughly once a month on average, always at the worst possible moment during a client review call. Since deploying the semantic layer, those discrepancies have dropped to zero in the eight months since we went live, because there is now only one version of every metric definition and everything pulls from it. We save around ten hours per week after using it. It also lets our engineering team focus on the core platform rather than maintaining a reporting layer, and powers everything client-facing without us building an analytics engine from scratch internally. Review collected by and hosted on G2.com.
The thing I'd most like to see improved is performance with large datasets. Around 90% of users say they perform slowly when processing large amounts of data for bigger reports, and they end up forced to export the raw data as a CSV file to use in another application. I hit this directly when one of our enterprise clients asked for a dashboard covering three years of transaction history across multiple product lines. The load time was long enough that the client mentioned it during our review call, which was not the impression we wanted to make. The workaround we currently use is pre-aggregating heavy datasets before they hit GoodData.AI, rather than querying raw data for large date ranges. That has improved load time noticeably, but it adds engineering work that shouldn't be necessary on a platform positioned for enterprise scale. GoodData.AI should prioritize native query optimization for large historical datasets rather than leaving users to architect their own runs outside the platform. Review collected by and hosted on G2.com.
Thank you for this detailed review! It's great to hear the semantic layer eliminated those monthly discrepancies and is saving your team around ten hours a week-exactly the outcome we aim for with governed metric definitions.
We hear you clearly on large dataset performance, this kind of specific example is exactly what helps our product team prioritize the right fixes.
Reach out to me at anna.slabunova@gooddata.com if you'd like to share more.
See how GoodData.AI improved


