How Alibaba Built a Cloud‑Native HR System That Cut Costs 100× and Boosted Speed 6×

Human Resource Management Systems (HRMS) centralize data for organization, recruitment, payroll, performance, and approvals. Globally recognized solutions include Oracle PeopleSoft, SAP, and Workday, while Chinese vendors such as Kingdee and Yonyou have long experience in eHR.

Alibaba adopted Oracle PeopleSoft HCM in 2006, which supported rapid growth for a decade. As the Alibaba ecosystem expanded to tens of thousands of employees across 24 countries, the legacy system could no longer meet functional, user‑experience, and integration requirements, prompting a move to a self‑developed solution.

Four major challenges of building a new system:

Complex migration of numerous tightly coupled modules (personnel data, performance, salary, reporting, permissions) with ten‑year historical data that must be cleaned and transferred without error.

Lack of reference implementations for large‑scale enterprises; Alibaba needed over 70 functional modules beyond typical ERP replacements.

Rapid ecosystem growth and internationalization demanded a platform that supports diverse business units and global requirements.

Cost control: the new system had to be maintainable with limited resources while delivering high business value.

To address these, the eHR team adopted a phased, incremental migration strategy.

Phase 1 – Design & Proof of Concept

Senior compensation and product experts formed a task force, selecting Python/Groovy for scripting, and designing the calculation framework, cloud task scheduling, and data encryption.

Phase 2 – Compute Framework Development

The team chose Groovy scripts running on Alibaba Cloud's big‑data service MaxCompute because Groovy integrates seamlessly with Java and is easy for Java engineers to adopt. MaxCompute enables efficient analysis of massive data, dramatically improving payroll calculation speed and security.

The system architecture separates business language, system language, and executable Groovy scripts, allowing users to edit high‑level business rules that are automatically translated into executable code.

Phase 3 – Payroll Logic Development

Three major releases covered intern salaries, equity taxation, and full‑time employee payroll, handling over 300 distinct calculation projects and complex scenarios such as leave adjustments, travel expenses, mid‑month hires, loans, benefits, social security base changes, and equity tax.

Phase 4 – Parallel Validation

From September 2016, an eight‑month dual‑system run rebuilt six months of historical data and validated year‑end bonus calculations across both systems.

The rebuilt eHR platform now supports over 40 products, including resume screening, learning, and equity management, with a small development team delivering rapid releases.

In May of the reporting year, Alibaba decommissioned all PeopleSoft HCM interfaces, fully transitioning to the self‑developed system, which now serves employees in 24 countries via web and mobile.

Key outcomes:

Using MaxCompute, payroll for tens of thousands of employees completes in 30 minutes versus 3.5 hours on PeopleSoft, a >6× speed gain that scales with data volume.

Annual licensing fees saved millions of dollars; overall maintenance costs dropped by two orders of magnitude.

End‑to‑end visualization of calculation processes makes the system easy to maintain and understand for non‑technical business users.

Platform‑centric architecture abstracts business logic into configurable Groovy scripts, enabling rapid addition of new scenarios such as 13‑month payroll, equity vesting, and tax automation.

Sensitive data is encrypted with Alibaba KeyCenter; even DBAs cannot view plaintext, enhancing employee privacy.

Mobile and PC portals provide real‑time access to salary, bonuses, equity, social security, and tax information, improving employee experience dramatically.

Looking ahead, Alibaba plans to evolve the HCM system into an intelligent platform covering globalization, ecosystem integration, mobility, data‑driven insights, and service‑orientation, leveraging Alibaba’s big‑data, AI, and cloud capabilities.