Case Study: Merkle reduces model runtimes and cloud costs with Qubole

Merkle cuts model runtimes in half with Qubole

Merkle, a global data-driven marketing agency, faced a challenge with its on-premises environment struggling to efficiently scale for machine learning processes. Their models, which involved processing millions of records, took about a day to run, causing projects to last up to three weeks. To modernize its operations, improve scalability, and reduce data-processing costs, Merkle sought a cloud-based vendor and selected Qubole.

By implementing Qubole's cloud-native data management platform, Merkle significantly reduced model runtimes from a day to five or six hours and shortened project durations from three weeks to just four days. The solution provided by Qubole enabled Merkle to test a wider range of models and scenarios quickly, improved efficiency for reporting, and reduced compute costs by multiple factors for the same workloads. Clients using Merkle's models saw an average 25% lift in sales and a 26% increase in revenue.

Merkle