A monolith that could not scale
The application, content library and admin tooling all lived in one Rails app. A change anywhere meant redeploying everything, and release confidence was low.
Case study
Legacy to modern stack migration
How we rebuilt a Rails monolith into a fast, searchable Next.js platform the Housebuild team can ship on every week.
Next.jsReactNode.jsTypesenseAWSMongoDB
- Industry
- Construction & building compliance
- Engagement
- 2023 — 2025
- Region
- Ireland · UK
- Stack
- Next.js · React · Node.js
At a glance
A platform rebuilt around search, compliance and speed — without a day of downtime.
0%
Growth in organic traffic
0
Lighthouse performance score, up from 58
0.0x
Increase in monthly active users
0%
Reduction in infrastructure spend
Scope of work
- Platform re-architecture
- Database transformation
- Search & relevance
- Technical SEO
- Native mobile apps
The brief
Client & project overview
Where the business started, what was holding it back, and the shape of the engagement.
Housebuild helps builders, architects and DIY renovators make sense of building regulations and find compliant products. The platform grew quickly on a Rails monolith backed by a single Postgres instance, and both were starting to creak under the weight of the document library.
Pages took seconds to become usable, search relevance was unpredictable, and every new regulation or product category meant a risky migration. Organic search drove most of the traffic, which made the slowdown expensive rather than merely annoying.
We re-platformed the product onto Next.js, with a search service built for structured regulation data and a document model designed to grow. The migration shipped behind a phased rollout, so customers never lost access while we moved them across.
Homepage & regulation search
The problem
What was standing in the way.
The constraints we found at the start of the engagement, and why the existing setup could no longer carry the business.
Postgres at its limit
Regulations, products and user data shared a single relational schema. Complex queries were slowing down and every content update risked a migration.
Search that got expensive before it got good
Algolia costs grew with the document library while relevance got worse. Regulation content is highly structured, and the ranking model could not reflect that.
Pricing that punished growth
Dyno-based hosting meant every traffic spike translated straight into cost. There was little room to tune infrastructure or add the services the product needed next.
SEO left on the table
Slow, client-rendered pages and incomplete metadata limited rankings on exactly the regulation questions customers were already searching for.
Nothing usable on site
The site was responsive but not practical on a building site. Teams needed checklists and regulation lookups on a phone, often with poor signal.
What we built
Our solution
We designed a phased migration that moved the platform without interrupting the business. We started with a full audit of the existing architecture, ranked the risks, and then rebuilt the critical path — search, content and the checklist engine — first.
Frontend
We moved the Rails views to a modern Next.js application, using hybrid rendering where it pays off:
- Server Components for content pages, so regulations render fast and index cleanly
- A shared component library used by the web app, admin panel and mobile apps
- Core Web Vitals pass on every template, with Lighthouse above 90
- Tailwind CSS design system for consistent, responsive UI across the product
Key improvementLargest Contentful Paint cut by 71% on mobile
Under the hood
Technology stack
The tools we chose, and why. Every choice is made for maintainability first, novelty second.
Next.js
React framework for hybrid static & server rendering
React
UI component library for building interfaces
Tailwind CSS
Utility-first CSS framework
TypeScript
Typed JavaScript for safer code
React Native
Cross-platform mobile app framework
The outcome
Results & impact
Measured against the baseline we captured before the rebuild. Where a figure is an estimate, we say so.
Technical achievements
Lighthouse score 58 → 94
Core Web Vitals now pass on every template, on desktop and mobile
Search cost cut by 71%
Typesense replaced Algolia, improving relevance and removing per-search pricing
Mobile apps shipped in 14 weeks
Native iOS and Android apps sharing business logic with the web app
99.99% uptime
Zero-downtime deploys with autoscaling and auto-healing in place
Business impact
4.1x monthly active users
Active user base grew 4.1x in the first five quarters after relaunch
312% organic traffic
Search traffic more than tripled as regulation pages began ranking
2.4x session duration
People stay longer and bounce less now that search actually helps
183% more conversions
Trial-to-paid conversion up 183% against the pre-migration baseline
In their words
“Intosoft played a vital role in building our MVP, showing strong understanding of both tech and business needs. They adapted quickly to feedback and supported us with responsive updates and improvements.”
Eoghain Ryan
CEO, Housebuild
Let's find what AI can do for your business.
Start with an AI Opportunity Audit, two weeks, a ranked map of where AI pays off, and an honest answer about what's worth building. No pitch decks, no pressure.
Get Started