14 years. One codebase. Zero rewrites.

Chapter 01 — The Problem

A travel platform that outlived three technology cycles.

ShermansTravel launched in the early internet era as a curated travel deals publisher. By the time Runtime Revolution came on board, the platform was handling millions in annual bookings, had accumulated years of business logic, and was facing a familiar pressure: modernize without breaking what works.

The instinct at companies like this is usually to rewrite. Start fresh, shed the technical debt, get back to a clean architecture. We've seen what happens to those rewrites. Most of them don't ship. The ones that do ship usually cost three times the estimate and lose half the business logic along the way.

We pushed back on the rewrite. We made the case for evolution — careful, incremental improvement of the existing codebase — backed by 18 years of doing exactly that.

Chapter 02 — The Approach

Evolve, not replace.

The first thing we did was understand the system: not just the code, but the business rules embedded in the code. Years of edge cases, special logic for deal types, affiliate tracking nuances, SEO conventions that drove real traffic. A rewrite loses this. An evolution preserves it while improving the architecture around it.

We introduced modern tooling incrementally — improving test coverage before refactoring, adding Elasticsearch to replace slow database queries for deal search, progressively modernizing the frontend without a big-bang migration.

Crucially: we kept the site live throughout. Zero scheduled downtime for major architectural changes. All improvements shipped as iterations, not as moonshot releases.

Chapter 03 — The Results

14 years. One codebase. Zero rewrites.

14 yrs

The ShermansTravel partnership is Runtime's longest. Fourteen years of continuous improvement on a single codebase. The platform handles millions in booking revenue annually on infrastructure that has been modernized piece by piece rather than replaced all at once.

The business logic accumulated over a decade is still intact — because we never threw it away. The performance improvements, the search modernization, the frontend updates — all shipped incrementally, all in production, all traceable.

This is what 18 years of production experience actually means: not the ability to build fast on greenfield projects (everyone can do that), but the discipline to improve existing systems without breaking them.