STALLION
MATCH.

When G1 Racesoft brought us in, Stallion Match was three years into development with another vendor and still didn't work.

The contracting team was a large software shop running the build as a waterfall delivery — fixed scope, sequential phases, monthly status reports. Which is exactly the wrong shape for a startup trying to find product-market fit.

But the structural problem wasn't the worst of it. The worst of it was that the team building a horse breeding analytics platform didn't know the difference between a stallion and a mare.

You can't model an industry you don't understand. And it showed in everything: business logic written as stored procedures inside the database, a frontend riddled with bugs, an over-licensed MSSQL setup quietly burning money, and algorithms that should have been backend code living as procedural SQL nobody else could maintain.

We inherited a codebase in genuinely bad shape.

The first month: learn the domain

Before we touched the code, we spent a month learning horse breeding. Pedigree structures. Bloodline analytics. Stallion-mare matching logic. What farm operators across Australia, New Zealand, Europe and Central Asia actually do day-to-day. The vocabulary. The data sources. The decisions the platform was supposed to support.

This sounds obvious. Most agencies skip it. We don't.

You can't build a tool for a domain you don't understand — and the previous vendor's three years of output was the proof.

The next six months: ship it

Once we understood the domain, we rebuilt.

• — Migrated business logic out of stored procedures and into a proper backend
• — Replaced the over-licensed MSSQL setup with PostgreSQL, dropping infrastructure cost significantly
• — Rewrote the frontend in ReactJS / NextJS — fast, stable, mobile-ready
• — Hardened security around what is genuinely sensitive commercial data for breeders
• — Optimized the algorithms instead of throwing more database licenses at the problem

We launched Stallion Match to the public six months after taking over.

For context: the previous team had three years and didn't ship a working product.

We didn't hand it off after launch. We became G1 Racesoft's long-term technical partner — and four years in, the product keeps getting smarter.

The most recent layer is AI. Stallion Match now runs AI-driven workflows across the matching and analytics surface — pulling patterns out of pedigree data, surfacing matches that the older rule-based system would have missed, and turning what used to be hours of breeder research into something that takes minutes.

Alongside the AI work, we've continued the unglamorous engineering that keeps a platform healthy:

• — Brought infrastructure costs down further as the user base grew
• — Layered in SEO infrastructure to support inbound growth across international markets
• — Mined and structured external breeding data to feed the analytics engine
• — Kept iterating on the product alongside the founders, week after week

The platform we shipped in 2022 isn't the platform that's running today. That's the point of having a partner who stays in the codebase.

Stallion Match is now live across four continents, in active use, and still growing.

Big dev shops aren't always the safe choice. Sometimes they're the expensive one.

A startup product needs a team that learns the domain before writing the code. That makes infrastructure decisions based on what the business actually needs, not what's easiest to templatize. And that stays in the codebase long enough to keep making it better — including layering in the AI work the product needs now.

That's the difference between a vendor and a partner.