Case Study: Yodle, a web.com company, accelerates BI analytics and triples product engagement with Elastic (Elasticsearch)

Powering a BI Application with Elasticsearch

Yodle, a web.com company serving over 50,000 local businesses, needed a fast, interactive BI product to combine website traffic, SEM, phone-call and review data. Their legacy Centermark batch stack (PostgreSQL, S3 Buckets, EMR/Scalding, HBase and CM v1 report services) struggled with requirements to backfill and re-backfill historical data, ingest fresh daily data, handle changing historical records, and deliver low-latency aggregations across many microservices and environments.

They rebuilt the pipeline as streaming jobs: Data Pipeline Service → Kafka → Storm analytics preprocessor → Elasticsearch powering the Report Querying Microservice and Angular analytics UI. The production ES deployment is a 5-node m4.2xlarge cluster (ES 2.4.1, doc values), indexing ~1TB of data (1.6B docs, largest index ~400GB; replicas → ~3TB). The new stack supports repeatable backfills and reprocessing, returns queries in ~2–3 seconds across 2K locations spanning multiple years, and helped drive ~3x product engagement.

Yodle

Mark Drago

Director of Engineering