CUSTOM SOFTWARE CASE STUDIES
Two Houston Legacy Replacements — Real Numbers, Real Timelines
Bottom Line Up Front (BLUF)
Two Houston companies — a commercial property management firm and a general contractor — were running critical operations on fragile legacy tools: a 15-year-old Microsoft Access database and a $81K/year fleet-tracking SaaS. We replaced both with custom-built systems using the same engineering methodology: extract the data, wrap the old interface, build the new one, cut over with zero downtime. Total engineering cost across both projects: under $60,000. Combined annual savings: $95,000+.
Case Study 1: Replacing a 15-Year-Old Access Database
It's an open secret in Houston commercial real estate: portfolios worth hundreds of millions of dollars are tracked on a single Microsoft Access file sitting on a local network drive. It works — until someone locks the file, the network blips, and the database corrupts.
The Problem
Access was designed for small-scale desktop use. When pushed to handle multi-user property management, lease tracking, and maintenance logs across multiple offices, it fails predictably. The symptoms are universal: the file takes minutes to open, simultaneous edits cause crashes, and running monthly financial reports freezes the workstation. The real cost isn't the crashes — it's the inability to integrate with any modern cloud API, permanently locking the firm out of automation.
| Metric | Before (Access) | After (Custom Web App) |
|---|---|---|
| Database load time | 2-4 minutes | Under 800ms |
| Concurrent users supported | 1 (file locking) | Unlimited |
| Monthly report generation | 45 min (manual export) | Real-time dashboard |
| Field access | Office desktop only | Any device, any location |
| Backup strategy | "Copy the file to a USB drive" | Automated daily with 30-day retention |
The Migration Playbook
Data Extraction
Automated schema translation lifted the raw tables from Access and migrated them to a secure PostgreSQL database hosted on AWS. Every record verified. Zero data loss.
The API Bridge
Instead of immediately rebuilding the interface, we configured the existing Access frontend to point to the new cloud database via ODBC. Staff kept using the interface they knew, but the data was now secure, backed up, and fast.
Modern Web Portal
With data safe in the cloud, we built a custom web application using React. Field agents could now update maintenance logs from iPads. Property managers could pull financial reports from home.
Phased Cutover
Features completed in the web portal, users transitioned off Access one department at a time. Dual-write phase ensured the original database remained intact as a fallback until full validation.
Timeline: 5 weeks. Total cost: $18,000. Downtime: Zero.
Case Study 2: Building a Fleet Tracker in 3 Weeks
A Houston general contractor was paying a fleet-tracking SaaS vendor $45 per truck, per month. At 150 vehicles, that's $81,000 per year — for a product where they used 10% of the features. They didn't need route optimization algorithms or fuel-tax reporting. They needed to know where their skid steers were. That's it.
The Build-vs-Rent Tipping Point
Off-the-shelf software makes financial sense when you're small. But as your fleet scales, renting software becomes an unjustifiable recurring tax. At 150+ vehicles, the math flipped decisively toward ownership.
The 21-Day Execution
Week 1: Infrastructure
Secure PostgreSQL database on AWS. Lightweight Node.js API to handle high-frequency geolocation telemetry from mobile devices. Architected for 500+ concurrent device connections.
Week 2: The Field App
Stripped-down React Native application installed on foremen's company iPads. Runs silently in the background, pinging GPS coordinates every 60 seconds without measurable battery drain.
Week 3: The Command Center
Astro 6 dashboard for the dispatch office. Single-screen interface mapping all 150 vehicles in real-time, color-coded by job site. No training required — the foreman looked at it once and started using it.
| Metric | SaaS Fleet Tracker | Custom Build |
|---|---|---|
| Year 1 cost | $81,000 | $15,000 (one-time build) |
| Year 2 cost | $81,000 | $1,200 (hosting only) |
| 5-year total cost | $405,000 | $19,800 |
| Features used | ~10% | 100% (built for their exact workflow) |
| Data ownership | Vendor-controlled | Firm owns everything |
ROI breakeven: Under 4 months. Annual savings: $79,800.
The Pattern: When Custom Software Beats SaaS
Both projects followed the same decision logic. Custom software makes financial sense when:
- You're paying for features you don't use — SaaS vendors build for the broadest possible market. If you use 10% of the features, you're subsidizing 90% of someone else's product roadmap.
- Per-seat licensing scales against you — at 150 seats, you're paying enterprise prices for a tool that treats you like a small customer.
- Data ownership matters — if the SaaS vendor goes under or gets acquired, your operational data is hostage. Custom systems put you in control.
- The workflow is simple but high-volume — complex, constantly evolving workflows favor SaaS. Simple, stable workflows with high transaction volume favor custom builds.
Read more about this framework in our Build vs. Buy Decision Guide or review our full execution log for additional project data.
Paying too much for software you barely use?
Get a Build-vs-Buy Analysis
We'll audit your highest-cost SaaS subscription and deliver a fixed-price comparison: what it costs to build vs. what you'll pay in licensing over 3 years.
Calculate Your Build vs. Buy ROI →