Beauty Retail: From Monolith to Microservice

Content Architecture That’s More Than Skin-Deep

Challenge

A premium beauty retailer was running personalization on top of Oracle ATG/BCC, a stack so rigid that every content update required an engineering dependency and personalization at scale was cost-prohibitive. With holiday peaks approaching and competitive pressure mounting, the business needed to exit the monolith and move to a headless CMS, DAM, and PIM stack, without disrupting the marketing calendar or losing data integrity across site, apps, email, and campaigns.

Outcomes

Content migration effort: Under 2 hours

Launch delivery: On-time before holiday peak

Editorial independence: Self-serve editorial

Unified content model: 496 atomic entries

Overview

Guided a premium beauty retailer’s leap from a legacy Oracle ATG/BCC monolith to a modern, microservices-based content ecosystem with Contentful at the core. The migration positioned the business for true omnichannel personalization at scale, while surfacing lessons about content modeling, change management, and SaaS governance.

Challenge

The retailer’s legacy stack included:

• Oracle ATG (Web Commerce): A powerful but rigid monolithic commerce engine, notorious for heavy deployments.
• Oracle BCC (Business Control Center): Content and merchandising tooling that slowed authors and required engineering dependency.

This setup made personalization prohibitively costly and slowed day-to-day content updates. The challenge: replace Oracle ATG/BCC with a headless CMS/DAM/PIM stack, while preparing workflows, governance, and tooling to handle the velocity and volume personalization demands.

One of the most complex examples: the homepage. With atomic content, each banner, promo, and slot can have multiple children (e.g., images, CTAs, text variants). By the time producers finished assembling the holiday homepage, the CMS showed 496 entries tied to that single page. This highlighted both the power and the complexity of atomic content modeling at enterprise scale.

My Role

Senior Technical Product Manager, Content Management Domain (CMS, DAM, PIM)

• Led RFP, vendor evaluation, and implementation for a headless CMS (Contentful).
• Designed the atomic content model and migration roadmap.
• Coordinated across creative, engineering, and marketing teams.
• Facilitated retrospectives to capture lessons on content modeling, training, and governance.

Process

1. Vendor Selection

• Ran a structured RFP with workshops to translate stakeholder pain points into measurable requirements.
• Narrowed to three headless CMS contenders, selected Contentful.

1. Content Model Transformation

• Shifted from template-bound BCC content to an atomic model (attributes → components → modules → views).

1. Migration Strategy

• Conducted a full audit of legacy ATG/BCC content.
• Built a category-by-category migration plan.
• Developed a custom migration utility (award-winning) to cut migration effort from hundreds of hours (manual) to under two (programmatic).

1. Automation & Tools

• AEM DAM integration automated asset selection.
• Extended Contentful with custom apps for editorial ease: including a Bulk Editor app allowing content producers to batch upload via CSV.
• Worked with RWS vendor to custom build touchless translation automation.

1. Omnichannel Enablement

• Designed a unified model powering site, apps, email, and campaigns.
• Ensured parity across devices and channels.

1. Retro & Stabilization

• Captured what went well: on-time launch before holidays, smooth rollout, strong teamwork, producers loved Contentful.
• Captured challenges: preview stability, rushed DAM/RWS integration, late change management, and overlapping personalization.
• Defined action items: user-friendly, stable preview as a next priority; business plan for dedicated Design System POD; clarify SaaS DevOps ownership.

Key Outcomes

• Winner, Quarterly Simplification Award for developing a Contentful data migration utility that saved engineers and producers hundreds of hours migrating 500+ category and brand pages programmatically.
• On-time launch before holiday peak — critical for business.
• Unified content model — one source of truth powering site, apps, email, and campaigns.
• Positive editorial adoption — faster workflows, reduced reliance on engineering.
• Faster time-to-market — creative automation and reusable templates shortened launch cycles.
• Homepage complexity revealed — 496 atomic content entries managed successfully, showing both scalability and the need for ongoing governance.
• Cultural shift — teams learned that content modeling is not optional; it’s the backbone of personalization.

Reflection

This migration wasn’t just a technology upgrade — it was a cultural transformation. Editors became autonomous, engineering shifted to enablement, and leadership learned that change management and governance are as critical as code.

Launching personalization alongside migration stretched the team — in hindsight, staggering those efforts would have eased adoption. But the project proved that moving from monolith to modular is the only way to scale personalization without burning out design and production teams.

Content governance isn’t glamorous, but without it, personalization topples. This project showed that smart systems thinking makes beautiful design sustainable.

Key Questions

How do you migrate hundreds of pages from a legacy CMS to a headless platform?

Bex conducted a full audit of the legacy ATG/BCC content, built a category-by-category migration plan, and wrote the product requirements and roadmap for a custom programmatic utility that migrated 500+ pages in under two hours. The utility won a Quarterly Simplification Award for eliminating the manual migration effort estimated at hundreds of web producer hours.

What does a headless CMS content model look like at enterprise scale?

The atomic model Bex designed had attributes building into components, components into modules, and modules into views. By the time editors assembled the holiday homepage, 496 atomic content entries were under management on a single page, proving the model could handle real enterprise complexity.

How do you ensure editorial teams adopt a new CMS after launch?

Through early involvement in the content model design, iterative training, and change management that addressed adoption risks before go-live. On this project, producers were independently managing Contentful workflows post-launch with no engineering dependency for routine content updates.