Buy a SaaS tool when your process is common and an existing product already does the job. Build custom software when the way you work is why customers choose you, or when you're paying for several tools and still copying data between them. Most businesses do both: buy the standard parts, build the part that's unique.
SaaS (software as a service) means renting software by the month, like a CRM or a booking tool. Custom software is built for your business and owned by you. Neither is better in general. The right answer depends on how standard your process is, how many people use it, and how much the workarounds already cost you. Below: how to decide, three-year costs, signs you've outgrown a tool, four real custom builds, and the risks on each side.
Build or buy software: the short answer
Buy a SaaS tool when
Your process is standard
- Your process is common: CRM, invoicing, bookings, email
- An existing tool does the job with little or no workaround
- You need it running this month, not this quarter
- Your team is small, so per-user fees stay low
Build custom software when
Your process is the product
- The way you work is why customers choose you
- You pay for several tools and still copy data between them
- Staff spend hours a week working around the tool
- Per-user fees grow faster than the value you get
The cut
Be honest with the first question. Most businesses believe their process is unique, and most of it isn't: invoicing, email, and contact records work much the same everywhere. Usually one or two steps are genuinely different, and those are the only ones worth building.
Custom software vs SaaS cost over three years
SaaS is almost always cheaper in year one. Custom can win later, as users grow and workarounds pile up. A simple example with round numbers: a team of 20 on a tool at $50 per user per month, compared with a custom first version plus support.
| SaaS tool | Custom software | |
|---|---|---|
| Year one | $12,000 in subscriptions | $30,000–70,000 to build, plus support |
| Years two and three | $12,000 a year, more as the team grows | A care plan only, $400–1,800 a month |
| Hidden cost | Staff time on workarounds and copying data | Someone has to look after it |
| You end up with | A tool you rent | Software you own |
On these numbers alone, the SaaS tool stays cheaper, which is the honest point: custom rarely wins on subscription savings by itself. It wins when the tool can't do a step your business depends on, when several subscriptions and staff hours add up, or when the software itself is what you sell. Build prices are in what a web product costs.
How long each option takes to set up
Speed is one of the strongest reasons to buy. A SaaS tool can be running in days, though moving your data in and training the team usually takes a few weeks. Custom software takes longer: an internal tool typically takes 4–10 weeks with me, and a product your customers use 6–14 weeks, counted from an agreed scope.
That gap matters less than it looks when the tool doesn't fit. A SaaS tool that's live in a week but needs daily workarounds keeps costing time for as long as you use it. If your deadline is close, a sensible path is to buy now, note exactly where the tool falls short, and use those notes as the brief for a custom build. How long a build takes, and what slows it down, is in how long a first version takes.
Signs you've outgrown off-the-shelf software
Tools rarely fail all at once. They slowly get replaced by workarounds. Two or more of these usually mean it's time to look at custom, or at least at connecting what you have:
- Staff copy the same details between tools every day
- You pay for several tools to run what is really one process
- The team keeps a spreadsheet next to the tool you pay for
- The one step that matters most needs a manual workaround
- Customers notice the handoffs between your systems
- You pay for features nobody uses, and still miss the one you need
If you built your process on a no-code platform instead of a SaaS tool, the signs are similar but the move is different, as covered in when to leave no-code for custom.
Four businesses that built custom, and why
Spreadsheets to one system: a school
Hongiranaran registrations, attendance, fees, and parent communication across registers and spreadsheets. A custom platform built around the school's own routines now holds 1,000+ students and 90+ staff, and automates about an hour of daily admin. It cost $32,000 over ten weeks.
Several tools to one workflow
A marketing technology company ran campaigns with emails, landing pages, gifts, and postcards across separate tools, with nobody tracking the whole journey. A custom campaign platform, built in 14 weeks for $30,000, now runs it in one place: 100+ campaigns and 200K+ recipients. On a larger scale, a confidential B2B sales and gifting platform replaced a stack of separate sales tools with one workflow.
When the software is the business
An on-demand printing startup wanted printing to work like food delivery: order online, printed at the nearest branch, delivered fast. No tool did that, because it was the business itself. The custom platform now handles 400+ orders a day across 40+ branches.
Buy the standard parts, build the unique part
The best answer is usually both. Custom software doesn't mean building everything: payments, email, login, and maps are better bought as services and connected. What you build is the layer that does the job only your business does.
- Keep the tools that fit. Accounting, email, and standard CRM work rarely need replacing.
- Connect them first. A small integration that moves data automatically is often enough, and far cheaper than a new system.
- Build only the missing piece. The step no tool handles, connected to everything else.
This keeps the build small and the monthly bills sensible. The same thinking applies to online stores, covered in Shopify or a custom store.
How to move from a SaaS tool to custom software
The riskiest part of going custom isn't the build, it's the switch. A team that relies on a tool every day can't stop for a week while data moves. This order keeps the business running:
- Export everything early. Check the export covers all the records you need, including history and attachments.
- Run both side by side. A few people use the new system while everyone else stays on the old tool.
- Move one team or process at a time. Not the whole company on a Monday.
- Cancel the old tool last. Only after the final data import is checked and the team has stopped reaching for it.
Risks of SaaS and of custom software
SaaS: you rent, they decide
Prices rise, features change or disappear, and a vendor can be bought or shut down. Your data lives on their terms. Reduce the risk by checking export options before you commit and keeping regular exports of anything important. Before signing an annual plan, also check how much notice the vendor must give before raising prices, and whether your price is locked for the contract term.
Custom: you own it, you maintain it
Custom software needs looking after: updates, fixes, and hosting. If only one developer understands it, you depend on them. Reduce the risk by keeping code and accounts in your name, asking for written notes on how it works, and budgeting for support, as explained in what maintenance costs.
Questions to answer before you decide
Write down short answers to these. They usually make the decision obvious, and they make any quote you ask for much more accurate:
If the answers point to building, start with the smallest piece that removes the biggest workaround. What goes into that first release is covered in what to put in a first version. Or send me your answers, and I'll tell you whether to buy, connect, or build.
