Deliver sub-second, interactive analytics directly to your customers, at scale.
When analytics shifts from internal report to live product surface, four things change at once: the user experience, engagement, revenue, and the speed of decisions.
Dashboards load the moment users arrive. Every interaction feels like part of the product, not a report — with no spinner standing between a question and its answer.
Three teams using Apache Doris to serve sub-second analytics directly to their customers, at concurrency, on live data.
“Replacing Flink’s window computing with Doris improved development efficiency, adapted to dimension changes, and reduced computing resources.”
Real-time analytics for the JD.com search box: overall search traffic, online A/B test monitoring, and trending search word lists, all refined to SKU-level granularity for business analysts.
“Doris lets us deliver fresh business insights to merchants without waiting for offline reports.”
E-commerce and platform merchant analytics, covering order analysis, conversion analysis, and user behavior analysis.
“Apache Doris gives us interactive analytics on large-scale, high-frequency data with low latency.”
Finance, advertising, and marketing analytics platforms serving external users.
Customer-facing analytics is a different workload than internal BI. Here’s what the serving engine has to get right, and how Apache Doris answers each one.
Customer-facing apps need both fresh data and fast responses. New events must become queryable within seconds, and analytical queries must return in sub-second time for dashboards, embedded analytics, and in-product workflows.
Thousands of users may query data at the same time. The analytics engine must maintain low latency under heavy concurrent load, not just perform well in isolated benchmarks.
Customer-facing analytics often serves many tenants, users, or embedded applications from the same platform. The engine must isolate data, workloads, and resources so one tenant’s activity does not affect another tenant’s performance, security, or experience.
Data already lives in open lakehouse formats, object storage, and existing data lake architectures. The analytics engine must query it in place, combine it with real-time serving data, and deliver fresh insights without creating another data copy.
New data becomes queryable within seconds, while dashboards and in-product analytics stay fast and interactive. Users see the latest activity as it happens, not after the next batch cycle.
