The problem
The client sold online legal documents and operational tools to travel agents and other sellers of travel. The business worked. The platform did not. Everything ran as a set of plugins on WordPress, one feature bolted to the next, and the white-label and licensing plans the company was selling against had nowhere to live.
The services agreement put it in the company's own words: it had “reached the limitations of this Minimum Viable Product and urgently need to migrate to a SAAS solution.” That sentence is the whole engagement. A company that has outgrown its first build does not need a rewrite of its features. It needs the platform underneath them to become something it can sell more than once.
What we built
- The multi-tenant core. The scalable platform the white-label and licensing plans depended on, built to carry more than one brand on one codebase.
- The user model. Secure signup and login, and the migration of the existing user base across, so the people already paying did not start again.
- The AWS estate. EC2, security groups, Route 53, SES and RDS, stood up from nothing.
- The toolchain. Jira, Docker and GitHub, with deployment procedures written for development, staging and production, staged so the cost of the lower environments stayed small.
- The documents the company had never had. A product requirements document, the logic diagrams, the infrastructure and stack documentation, and a roadmap.
How we ran it
We provided and managed the Phase 1 team under our own paper: one front-end developer, two back-end developers, a part-time project manager, and systems support as it was needed. Frank Robles held the fractional CTO seat, which meant the architecture decisions and the scope arguments as well as the status report.
The first working deployment was targeted inside four months. A weekly call every Thursday morning carried the plan, and the environments were built in an order that kept the hosting bill down while the platform was still being assembled.
- AWS
- EC2
- RDS
- Route 53
- SES
- Docker
- GitHub
- Jira
- Multi-tenant architecture
What it produced
The platform work was delivered into the client's environment. The engagement ran from December 2022 to about May 2023 and ended when a successor agreement stalled, which is a commercial outcome rather than a technical one. Adoption and revenue figures belong to the client and are not ours to publish.
One piece came back to us. An e-signature component defined in that engagement is the direct ancestor of Sign42, the document product Div42 owns today.
What we can start on
A legacy build that has reached its limit and needs a platform underneath it. A multi-tenant rebuild with the user base carried across. Or an AWS estate, a deployment process and the written architecture for a team that has been shipping by hand.