Small-business owners ask us a version of this question constantly: should I just buy [some SaaS], or should you build me something custom?
The honest answer is almost always: buy the SaaS, and accept its flaws. But sometimes custom is the right call. Here's the framework we use to decide.
The four questions
1. Does a mature SaaS exist that covers ≥80% of what you need?
Mature means: has been around 5+ years, has integrations with the other tools you pay for, has 10K+ paying customers. For almost every small-business problem — CRM, accounting, scheduling, email marketing, HR, inventory — a mature SaaS exists.
If yes → buy the SaaS, even if the remaining 20% is annoying. Usually the right answer.
If no → go to question 2.
2. Is the missing 20% actually load-bearing for your business?
Be honest here. Most "missing features" are nice-to-haves. The actual load-bearing part is usually: we need this data to connect to that other tool. That's an integration problem, not a "build the whole thing custom" problem.
If the missing 20% is an integration → buy the SaaS + pay someone $3K-8K to build the connector. Still not a custom-build situation.
If the missing 20% is a genuine core feature that no SaaS has → go to question 3.
3. Will this problem still exist in 3 years?
Custom software is a 3-year commitment, minimum. You're signing up for: the initial build, bug fixes, feature additions, security patches, dependency updates, and the occasional rewrite when something fundamental changes.
If the problem might be solved by an incoming SaaS competitor in 18 months → wait. The cost of building-then-ripping-out is worse than waiting.
If the problem is structural to your business and will still be there in 3 years → go to question 4.
4. Is the value ≥ 10x the total 3-year cost?
Add up: build cost + annual maintenance (roughly 15-20% of build cost per year) × 3 years. Then compare to the value of the problem.
A rough table:
| Build size | 3-year total cost | Minimum annual value to justify |
|---|---|---|
| Small tool ($10K build) | ~$15K | $5K/yr minimum |
| Medium app ($50K build) | ~$80K | $25K/yr minimum |
| Full platform ($150K build) | ~$240K | $75K/yr minimum |
The 10x rule isn't arbitrary. Software engagements usually run over by 30-50%, and maintenance creeps. The 10x buffer protects you.
If the value is ≥ 10x → build it. You've passed all four gates.
If the value is < 10x → buy the SaaS, even if imperfect. The math doesn't work.
Worked example
A client approached us wanting to build a custom appointment-scheduling system because the SaaS options "didn't match our workflow."
Walking through the gates:
- Gate 1: Calendly, Acuity, and Setmore cover 85% of their needs. ✅
- Gate 2: The missing 15% was "book 2-hour slots with a 15-min buffer for specific staff." Acuity actually does this; the client just hadn't configured it. ❌
Total custom-build project avoided: $28K. We spent 90 minutes with the client configuring Acuity instead. They paid us $300 for the consultation. Both sides happy.
When to break the rules
There's one scenario where the framework doesn't apply: when the thing you're building is your actual product or competitive moat. If you're an e-commerce company and your custom pricing engine is what beats your competitors, build it custom no matter what. The framework is for operational software, not strategic software.
The uncomfortable truth
Most small-business owners don't need custom software. Most of the SaaS they already pay for, they're using at 40% of its capability. The single highest-ROI thing most of them could do is hire someone to actually configure the tools they already have.
That's not a pitch we make often because it doesn't pay well. But it's true, and we'll tell you when it applies to you.