CASE STUDY · NASDAQ-LISTED SAAS LEADER

$3M+ saved annually in customer-facing analytics

A NASDAQ-listed SaaS company with 68,000+ customers serves 10 million+ in-app dashboard queries a day on one lakehouse. TCO fell by ~60% once fully deployed, with zero-failure SLAs intact.

INDUSTRY · B2B SAASENVIRONMENT · AWS · DELTA · UNITY CATALOGUSE CASE · CUSTOMER-FACING ANALYTICS
SIX-WEEK SOAK TEST, REAL PRODUCTION TRAFFIC
METRIC
BEFORE
WITH E6DATA
Total cost of ownership
2x cost jump every 18 months
~60% lower, $3M annualized
p95 latency, web dashboards
30s in early testing
1.5-2s at 60 QPS peak
Failure SLAs on schedules and unloads
non-negotiable requirement
zero failures, met across workloads
Copies of data (web vs exports)
separate copies per workload
one lakehouse, web and non-web
"The moment e6data was able to meet the web SLAs, it became a very easy decision to make one data lake that serves both web and non-web, using e6data as a query engine."
ENGINEERING MANAGER · NASDAQ-LISTED SAAS COMPANY
THE STORY

One lake for web and non-web

The bake-off ran e6data against three other engines on real production traffic. The web SLA was the bar; the bill was the tiebreaker.

01

The stakes

In-app dashboards serving 10 million+ queries a day, on highly normalized data with 10+ table joins per query, against a sub-2-second p95 requirement. Customer-facing analytics is the product, not a feature.

02

The wall

Compute cost doubled every 18 months on the incumbent engine. The team also wanted a lakehouse architecture for AI and ML use cases, interoperability, and an exit from vendor lock-in, without giving up functionality.

03

The evaluation

Before committing, the team set the bar: the criteria that would decide it - p95 latency under real concurrency, cost at scale, security and private-link connectivity, and ease of adoption - measured on their own data rather than synthetic benchmarks.

04

The bake-off

Four engines, five criteria: performance, zero-failure SLAs, cost, ease of adoption, and security including private link connectivity. Evaluation started on data exports, then replayed tens of millions of production queries against the dashboard workload.

05

The result

p95 went from 30 seconds in early testing to 1.5 to 2 seconds at 60 QPS peak, with every failure SLA met. TCO fell ~60%, worth $3M a year fully deployed, and one lakehouse now serves web dashboards and exports from the same tables.

Book a demo on your own workloads

Reach out to book a demo, share challenges you're facing, and tell us how this fits into what you're currently working on or thinking about.

Prefer to self-serve? Problems we're solving

An actual person replies. By submitting, you acknowledge your personal information will be processed in accordance with our Privacy Policy.