The five questions
1. Fit: how much do you work around the tool?
List the workarounds your team does every week: spreadsheets on the side, copying data between systems, manual steps the software can't handle. If the workarounds cost hours per person per week and come from the tool not matching your process, that's a build signal.
2. Three-year cost: what will each option really cost?
For SaaS, include seats at your expected headcount, add-ons, implementation, and admin time. For custom, include the build, 15–20% a year for maintenance, and hosting. Year one often favors SaaS; year three often doesn't.
3. Integration: how many systems need the same data?
When a sale must become a job, a job an invoice, and an invoice a payment - across four tools - the cost of re-typing and reconciling can dwarf licence fees. Custom software (or custom integrations) is often the fix.
4. Risk: what happens if it goes wrong?
Buying is lower delivery risk; building carries project risk that a good partner reduces with fixed scopes, weekly demos and phased releases. On the other side, relying on a vendor carries pricing, roadmap and lock-in risk.
5. Ownership: does owning the software matter?
If the software embodies how you win - your pricing logic, your service model, your data - owning it can be a strategic asset rather than a cost.
A decision table
| Situation | Usually the right move |
|---|---|
| Standard process (payroll, accounting, basic email marketing) | Buy |
| Small team, standard sales pipeline | Buy a SaaS CRM |
| Internal tool, few users, changing requirements | Low-code or no-code |
| Customer-facing portal with your branding and workflows | Build |
| Several tools that each do part of the job | Build the integration layer, or a custom system that replaces the glue |
| Per-seat costs rising faster than the value delivered | Price a custom build |
| Core process that is how you compete | Build |
What about no-code and low-code?
Tools like Retool, Bubble, Airtable and Softr are excellent for internal prototypes and simple tools. They get harder when you need customer-facing polish, complex permissions, high data volumes, or when per-user pricing scales with your headcount. A common path is to prototype in low-code, prove the workflow, then build the durable version.
Evidence that companies are building more
Retool's 2026 Build vs. Buy report, a survey of 817 people who are Retool customers or builders (so not a neutral sample), found that 35% had already replaced at least one SaaS tool with custom software and 78% expected to build more internal tools. Take the exact numbers with caution, but the direction matches what we see: modern tooling has made focused custom software cheaper to build and maintain.
The hybrid answer
The best answer is often "buy the platform, build around it":
- Keep your SaaS CRM, and build a custom operations system connected to it.
- Keep your accounting software, and build the job costing it can't do.
- Keep your helpdesk, and add a custom AI assistant that works on WhatsApp.
Next step
If the five questions point toward building, the next step is a precise scope and a fixed price. See our custom software cost guide, or talk to us about custom software development - we'll tell you if buying is the better call.
Sources
Prices, plans and laws change. We review this guide periodically; check the linked sources for the latest figures. Nothing here is legal or financial advice.
