One Shared CMS and Serverless Platform for a Group of Companies
Each group company ran its own site, blog, events and forms separately, so shared changes were made several times and cost scaled with site count. We built one CMS on AWS serverless, and each company keeps its own identity.
- Service
- Software Solutions
- Industry
- Enterprise IT & Cloud
- Region
- Japan
- Engagement
- Lab-type ODC (dedicated team)
Outcomes
- One CMS for every group site
- Usage-based serverless cost
- Brand identity preserved
Stack
Headless CMS · AWS serverless · CI/CD · Infrastructure as code
Written byOriental Nippon Team
A group of related companies in Japan each had their own web presence: a site, a blog, event announcements, and contact forms. Each had been built and maintained on its own terms.
So a change to a shared pattern had to be made several times, publishing an event meant knowing which system that company used, and enquiries landed in different places.
The problem
- Repeated work. One change, made once per site.
- Different systems. Publishing depended on knowing each company's setup.
- Scattered enquiries. Contact forms fed different inboxes and were handled inconsistently.
- Cost by site count. Infrastructure cost grew with the number of sites, not with traffic.
| Before | After | |
|---|---|---|
| Content systems | One per company | One shared CMS |
| Publishing | Different per company | Same pipeline for all |
| Contact forms | Routed differently | Central routing |
| Hosting cost | Per site | Tracks actual usage, AWS serverless |
What we built
A shared CMS
Sites, blogs, events and forms for the whole group, with each company keeping its own identity.
Shared delivery pipelines
Publishing and deployment work the same way for every company.
Serverless infrastructure on AWS
Cost follows traffic, not the number of sites.
Infrastructure as code
The whole setup defined in the group's own repository.
Technology and the role it played
| Technology | Role in this project |
|---|---|
| Shared CMS | One content model for every company's site, blog, events and forms |
| AWS serverless | Hosting whose cost follows traffic |
| CI/CD pipelines | The same publishing path for every company |
| Infrastructure as code | The whole setup defined in the group's repository |
How we delivered it
We audited every site first to separate real differences from accidental ones. Most of what looked different was not.
- Audit the estate. What each company runs, and what is genuinely different.
- Design the shared model. Around real differences, not accidental ones.
- Build platform and pipelines. With one company as the pilot.
- Migrate site by site. Each with its own content and switch-over.
- Hand over publishing. To each company's own team.
01Estate audit
2 wks
What each company runs; real versus accidental differences
02Shared model design
2 wks
Content model, identity handling, form routing
03Platform and pipelines
4–6 wks
CMS, delivery pipelines, serverless infrastructure
04Pilot company
2–3 wks
One site migrated and published from the new platform
05Migration
ongoing
Site by site
Risks and how we handled them
| Risk | How we handled it |
|---|---|
| Brands flattened into one look | Shared model designed around real differences, keeping each identity |
| Migrating every site at once | One pilot company, then site by site |
| Editors dependent on developers | Publishing handed over to each company's team |
Outcome
Websites, content, events and forms are managed centrally across the group, on infrastructure whose cost follows usage.
Each company still looks like itself. A shared platform that flattens brand identity tends to get abandoned by whichever company cares most about theirs.
What the client keeps
- The platform and its source
- Infrastructure definitions, in their own repository
- Delivery pipelines
- The content model
- Per-company publishing documentation
Sources and further reading
- Serverless on AWS, AWS
- AWS Well-Architected Framework, AWS
- The Twelve-Factor App, 12factor.net
Facing something similar? This is the kind of work our Software Engineering & Cloud team delivers every day. Talk to us about your stack.
