Custom System vs. SaaS: A 3-Year Cost Breakdown for SMEs
SaaS subscriptions look affordable per month, but three years can cost over HK$700k. Custom systems require upfront investment but save money long-term. Here's how to pick the right path for your business.
SaaS subscriptions look tempting at first glance: free trials, someone else handles maintenance, cancel anytime. Yet many Hong Kong SME owners never calculate what three years of SaaS really costs—or whether building a custom system would actually save money.
SaaS pricing is deceptively high over three years
Picture a 50-person company using a CRM SaaS at US$50 per user per month. That's 50 × US$50 = US$2,500 monthly (roughly HK$19.5k), US$30,000 yearly (HK$234k), and US$90,000 over three years (HK$700k). Add setup fees, consultation hours, staff training, and custom API integrations with your other systems. Every time your workflow doesn't fit the SaaS template, you'll hear "that requires our Professional tier" or "not supported—build it yourself." A conservative estimate puts real three-year spend at HK$800k+.
Vendor lock-in is the hidden killer: once dependent on a SaaS platform, switching costs explode. Data migration, retraining, rebuilding lost workflows—the price of escape becomes prohibitive. If the vendor changes pricing or shuts down, you have no say.
Custom systems carry upfront cost but lower long-term expenses
Building a system in-house demands higher initial investment. A mid-complexity system (order management + accounting integration) typically costs HK$150,000–250,000 and takes 3–6 months. Maintenance runs HK$3,000–8,000 monthly after launch. Three-year math: HK$200k (dev) + HK$5k (monthly) × 36 = HK$380k—nearly 50% cheaper than SaaS, and you own everything.
Ownership means control. Your workflows evolve on your terms, not at the mercy of vendor updates. When staff leaves, your codebase and domain knowledge stay in-house. The next developer steps in without rebuilding from scratch.
Pick the right path: depends on scale and workflow complexity
No one-size-fits-all answer. Under 5 people with standard workflows (sales, invoicing, inventory)? SaaS wins: zero maintenance overhead, rapid deployment, predictable costs. Between 10–50 people with custom processes (e.g., a clinic linking appointments to staff scheduling)? Custom systems pay for themselves in 18–24 months. Over 50 people across multiple departments with diverse needs? A hybrid approach wins: one core SaaS (CRM, accounting) plus custom modules (reporting, automation) built separately.
Three real Hong Kong examples: a 60-person restaurant group using POS + custom accounting integration saves HK$150k/year. A Central clinic's custom booking + staff scheduling system paid for itself in 18 months. A 50-person import-export trader runs SaaS CRM alongside a custom reporting engine and gained 10× more system flexibility.
Five common mistakes to avoid
- Myth 1: SaaS has zero maintenance burden. False. Plugin conflicts, upgrade breakage, API changes all disrupt your workflow. Maintenance costs are hidden, buried in employee overtime.
- Myth 2: Once a custom system is built, costs stop. False. Every software needs bug fixes, security patches, and business logic updates. Monthly fees aren't optional.
- Myth 3: A freelancer can quickly build "good enough." Risky. No documentation, no git repository, vanished developers—a different kind of lock-in.
- Myth 4: SaaS is scalable, custom systems are fragile. Depends on who built it. A well-architected custom system scales. A bloated SaaS subscription suffocates under growth.
- Myth 5: Always choose the cheapest option now. Total cost of ownership over three years matters far more than monthly spend.
Three-year cost breakdown: side-by-side
| Option | Initial | 36 Months Fees | Hidden/Switching | 3-Year Total |
|---|---|---|---|---|
| SaaS (50 users) | HK$50k | HK$700k | HK$50k | HK$800k |
| Custom System | HK$200k | HK$180k | HK$0 | HK$380k |
| Hybrid (SaaS + custom) | HK$100k | HK$350k | HK$20k | HK$470k |
Five questions to ask before signing
If you choose SaaS:
1. How long is the contract lock-in, and what's the cancellation penalty? Some SaaS vendors demand 12-month commitments with exit fees. Non-negotiable terms are red flags.
2. Can you export your data and access via API without restrictions? Your data must be portable. Never accept "you can export, but only in our custom format."
3. Which custom features cost extra, and how much? Get it in writing. "Professional tier pricing TBD" means surprise bills later.
If you choose custom:
4. Who owns the source code and database? Must be your company, never the developer. Insist on it in the contract.
5. What happens if the developer stops supporting the system? Define maintenance duration, handoff procedures, and documentation requirements. Require source code escrow if it's a small firm.
FAQs
When should I migrate from SaaS to custom?
When annual SaaS cost exceeds 36 months of custom maintenance fees. Example: SaaS costs HK$300k/year, custom maintenance is HK$5k/month (HK$180k for 36 months). Custom is cheaper. But factor in migration cost (data transfer, staff retraining)—breakeven usually arrives by year two or three of operations.
What if the developer who built my custom system disappears?
Don't hire solo freelancers. Use firms with team backup. Require source code escrow (code held by a third party if the developer fails), complete documentation, and code review processes. Pay more for a reputable firm if needed. You'll sleep better.
How do I integrate SaaS and custom systems?
Three methods exist: one, native SaaS API (most stable); two, automation tools like Zapier or Make (cheaper but limited); three, scheduled data export-import (clunky but bulletproof). Pick based on your technical capacity and how often data needs to sync.
I bought SaaS but regret it. Can I escape?
Yes. If the contract allows, migrate to custom or a different SaaS. Budget one to two months for transition. Data mapping is time-consuming but doable. Next time, negotiate a 30-day trial period and a contract with no lock-in clauses.
Software choice is never perfect—only right-sized. Spend time evaluating total cost of ownership, testing pilots, and negotiating terms upfront. It costs more than guessing, but costs far less than regret.