Every operations manager in Australia will eventually hit the moment where the spreadsheet stops working, the SaaS bill balloons, or the workflow you've stitched together across five tools starts breaking. That's the moment you ask: should we just build our own?
Sometimes the answer is yes. More often, it's no. Here's how we help clients work through it — five questions, a cost model, and two real client stories where the answer went different ways.
The five decision questions
1. Is your process a competitive differentiator?
If the way you do something is what makes your business better than competitors, custom software might be worth building. If your process is generic — HR admin, invoicing, project management — buy the SaaS. Nobody wins a market by having marginally better generic tooling.
Example: STA Consulting Engineers had a proprietary calculation methodology for waffle-raft compliance that no SaaS could replicate. Custom software was the right call. But their HR runs on Employment Hero and their invoicing on Xero. Nobody wins by rebuilding those.
2. How many users, how often?
SaaS pricing is usually per-seat or per-usage. Custom software is a fixed build cost plus hosting. The math flips at scale. If you have 50 users hitting a $50/user/month SaaS, that's AUD $30,000/year — probably cheaper to build custom over 3 years. At 5 users, definitely not.
3. Does the SaaS actually fit?
SaaS products are opinionated. They assume a particular workflow. If you're contorting your business to fit the tool, or maintaining a 40-page manual explaining the workarounds, the tool doesn't fit. Custom lets you build to your actual process — but only worth it if the process is stable enough that it won't be different in a year.
4. What's your integration surface?
Every additional tool means another integration to maintain. If your current stack is Salesforce + Xero + Slack + HubSpot + Notion + custom Google Sheets glue — that's five APIs, five auth systems, five sources of truth. Sometimes replacing all of it with one custom platform saves more than the build cost in maintenance alone.
5. Can you afford to be wrong?
SaaS: you can cancel next month. Custom: you've spent AUD $60k+ and it's yours forever. If your requirements are still fluid, buy SaaS first. Learn what you actually need, then build (or extend the SaaS with a lightweight custom layer) once you know.
The 5-year cost model
Rough comparison for a workflow tool serving 25 internal users. Real numbers from client discussions:
Option A: Mid-tier SaaS
- License: 25 × AUD $60/user/month = $1,500/month = $18,000/year
- Onboarding + config: $5,000 year 1
- Custom integrations: $8,000 year 1, $2,000/year maintenance
- Annual price increases: ~8%/year (industry standard)
- 5-year total: AUD $122,000
Option B: Custom build
- Initial build: AUD $80,000 (4-month engagement, 2-person team)
- Hosting: AUD $2,400/year (AWS or Cloudflare)
- Maintenance: 15% of build cost/year = $12,000/year
- Feature additions: budget $15,000/year for evolution
- 5-year total: AUD $196,000
On price alone, SaaS wins. But price alone isn't the story:
- Custom includes ~$60k of feature additions over 5 years (your product evolves)
- Custom retains its full value at year 5 (you own the codebase)
- SaaS renewal costs will keep rising indefinitely
- Custom eliminates 3–5 SaaS subscriptions if it's a workflow platform (typically another $30–50k of savings)
Reality: the break-even is usually 3–4 years for a well-scoped custom build. After that, custom starts pulling ahead — sometimes dramatically.
Case study 1: We said build (STA)
STA Consulting Engineers had a spreadsheet-based waffle-raft compliance calculator. Ten engineers used it. It was slow, error-prone, and the calculations couldn't be validated by juniors.
- Build cost: AUD $65,000
- What replaced: 4 spreadsheets, an Access DB, and 30 minutes per report of manual formatting
- Result: Report time down from 90 minutes to 12. Junior engineers can now run compliance checks with senior review. Payback: 7 months on time saved alone.
Why it worked: proprietary methodology, stable requirements, high-cost users (engineers billing $180/hr), 5+ year lifespan.
Case study 2: We said buy (a NZ non-profit)
A NZ non-profit came to us wanting a custom volunteer management system. Their existing tools: Google Forms + Sheets + a mailing list.
- Our recommendation: Better Impact ($USD 200/month) + a lightweight Zapier integration
- Alternative custom build: AUD $45,000 + $6k/year hosting/maintenance
- Result: Live in 3 weeks. Total year-one cost: $4,800. Custom would've been 10× that with no meaningful upside.
Why buy worked: standard workflow, small user count, tight budget, and the SaaS did 95% of what they needed.
The middle path
Many of the best decisions we help clients make aren't build-vs-buy — they're buy plus extend. Use the SaaS for the 90% of standard workflow. Build a small custom layer for the 10% that's specific to you.
Example: use HubSpot for the CRM, but build a custom quoting tool that integrates with it. Total cost: SaaS license + AUD $20–40k custom build. Best of both worlds.
Our default advice
Buy SaaS for the first 2 years while requirements stabilise. If, at year 2, you've genuinely outgrown it or are paying more than AUD $40k/year, revisit custom. Until then, focus your engineering budget on the thing that is your competitive edge — not on rebuilding tools that already exist.
Found this useful? Two new posts a month — back to all insights.