"Just use Notion / Airtable / Odoo" is great advice — until it isn't. Here's the actual decision framework, with numbers.
When SaaS wins (use it)
- Your workflow matches the tool's defaults out of the box. Like 80%+ match, not "we can make it work."
- You have fewer than ~50 users. SaaS pricing usually stays reasonable here.
- Your data isn't a competitive secret. Hosting your customer list on Airtable is fine. Hosting your proprietary trading algorithm is not.
- You don't need it to integrate deeply with 3+ other internal systems.
- The team using it can change their workflow to match the tool.
When custom wins (build it)
- You're paying every year for SaaS that doesn't quite fit, and you're working around it constantly.
- The workflow lives in your team's heads — and the SaaS is forcing them to change a process that actually works.
- Your data has compliance, residency, or competitive implications.
- You need real-time, deep integration with your existing systems (ERP, accounting, custom databases).
- You're hiring people to do work that should be automated, because the SaaS can't be customized to automate it.
The 3-year cost crossover
Most companies don't notice when they cross from "SaaS makes sense" to "we should have built this." Here's a rough math:
- SaaS path: subscription fees × 3 years + the team time spent working around the tool, every month.
- Custom path: the phase-1 build + yearly maintenance over the same 3 years — and you own the result.
The crossover is faster than most teams expect. By year 2, custom is already cheaper. By year 3, it's not close.
The trap: half-building inside SaaS
"We'll just add custom code to Airtable / a Zapier workflow / a Notion automation." This is how SaaS bills triple. You're paying SaaS premium + paying engineers to work around limitations.
If you're writing more than 200 lines of integration code inside or around a SaaS tool, you've already crossed the line — you just haven't admitted it.
What "custom built" actually looks like in 2026
- Built on Next.js + PostgreSQL + Cloudflare. You own the repo.
- Phase 1 ships in 4–8 weeks. Replaces your most painful workflow first.
- Phase 2 (3 months later) adds the next two workflows.
- By month 12 you've replaced 60–80% of your SaaS stack.
- Monthly retainer to evolve it costs less than the SaaS you cancelled.
Questions to ask before building
- How much are we paying for SaaS today that we're going to cut? (Be specific.)
- What workflows are we doing in Excel because the SaaS can't?
- If we don't build, what's the cost of staying as-is for 3 more years?
- Who owns the system after launch? (Should be: you, with a vendor on retainer for changes.)
Wondering if your situation crosses the line? Tell me what you're using today and I'll give you an honest answer — even if the answer is "stay on SaaS."