Build vs Buy Software: The Four-Option Decision Most Guides Miss
Why build vs buy software is really four options
- ✓
Buy. You license an off-the-shelf product and change your process to fit it.
- ✓
Integrate. You keep your current systems and connect them, so data moves once instead of being typed twice.
- ✓
Automate. You add workflow automation around existing tools: forms, approvals, reminders, scheduled reports.
- ✓
Build. You commission custom software because nothing on the market fits a process that genuinely sets you apart.
Fix the process before you choose any option
Automating or replacing a broken process just makes the mess faster. So before you compare vendors, map how the work actually flows today.
Start with the Thursday-afternoon report somebody builds by hand, then trace where the numbers come from. Look for the approval stuck in one person’s inbox. Find the spreadsheet that exists because two systems disagree. Often, half the “software problem” disappears once duplicate steps go and one team owns the data.
That’s why a business operations audit usually comes before any software build vs buy decision. After all, you can’t score process fit against a process nobody has written down. As a business automation and operations partner, we treat this as step zero, because technology is the tool, not the goal.
A decision table: which signals point to which option
Match the signals you see to the option they point towards. If your signals point in two directions, score both options using the method in the next section.
| Signal you’re seeing | Option | Why |
|---|---|---|
| Standard process (payroll, timesheets, basic CRM) | Buy | Vendors have solved it already |
| Data typed twice between two good systems | Integrate | The tools work; the gap doesn’t |
| Approvals and reminders chased by email | Automate | A workflow layer fixes it |
| Reports assembled by hand from exports | Integrate, then automate | Clean data first, then schedule the report |
| The process wins you work and no product fits | Build | Owning the code can pay back |
| Current system is unsupported or end-of-life | Buy or build | Compare replacement options honestly |
| Nobody internally can own a new system | Buy or automate | Custom software without an owner decays |
A quick reality check on that last row. Custom software needs a product owner long after launch. If nobody on your side has the time, building is a bad idea, however good the specification looks.
How to score the software build vs buy decision
- 01
Write down the redesigned process and the measurable outcome you want, such as hours saved per week.
- 02
Build a five-year total cost of ownership for each option. Include setup, licences, maintenance, hosting and the manual time that remains.
- 03
Score each option 1 to 5 on the six criteria, with at least two people scoring independently.
- 04
Multiply each score by its weight, add them up and divide by 100.
- 05
Then test the top two options with a small prototype or a vendor trial against real cases before you sign anything.
Worked example: five-year TCO for a subcontractor compliance process
- ✓
Buy a compliance SaaS tool at £18 per user per month for 40 users. That’s £8,640 a year plus £6,000 setup. However, details still get re-keyed into accounts, so 4 hours a week of manual work remain (£5,520 a year).
- ✓
Integrate the document module of the project platform they already pay for with the accounts system. Setup is £14,000, running costs are £2,400 a year, and 1 hour a week of manual work remains (£1,380 a year).
- ✓
Automate with workflow automation: a form, a reminder workflow and an expiry alert on top of existing tools. Setup is £9,000, tooling is £1,800 a year, and 2 hours a week remain (£2,760 a year).
- ✓
Build a custom compliance portal for £85,000. Maintenance runs at 15% a year from year two (£12,750), hosting at £1,200 a year, and half an hour a week remains (£690 a year).
When building custom software is the right call
Build when the process is a genuine competitive edge, no product fits it, and you have someone to own the result for years. All three conditions need to hold.
Even then, don’t commission the full system on belief. Test the riskiest assumptions with a prototype first. Our technology feasibility assessment exists for exactly this: we prototype, test against real use cases and tell you whether to build, buy or automate, including the option of not proceeding.
If you do build, the choice of supplier matters as much as the choice to build. For a Lithuanian client planning an AI-powered platform, we ran the needs analysis, wrote the RFP and led vendor selection for the customs knowledge platform, then scored every proposal on technical skill, project management and budget fit. If you’re at that stage, our guide on what an RFP is and how to run one walks through the process. Also check supplier security. The NCSC’s supply chain security guidance is a practical checklist for any vendor that will hold your data.
Common mistakes in the build vs buy decision
- ✓
Comparing licence fees instead of total cost. In the example above, an £8,640-a-year licence became £76,800 over five years once we added setup and the manual work it left behind.
- ✓
Buying to fix a process problem. A new platform on top of a broken approval chain gives you a broken approval chain with a new login.
- ✓
Ignoring the tools you already pay for. Many firms use a fraction of their current platforms. Check what you own before you shop, and read our guide to connecting business systems first.
- ✓
Building for pride. Custom software feels like an asset. For a commodity process, it’s a liability with a maintenance bill.
- ✓
Skipping the exit plan. If you can’t export your data in a usable format, the vendor owns your process.
Frequently asked questions
What does build vs buy software mean?
It's the decision between commissioning custom software and licensing an existing product. In practice, there are four options: buy, integrate the systems you already have, automate the work around them, or build. The right answer depends on five-year total cost, how well each option fits your redesigned process and whether the process gives you an edge.
Is custom software cheaper than off-the-shelf in the long run?
Rarely, for standard processes. Custom software carries the build cost plus a maintenance bill every year you run it (we used 15% of the build cost in the example above). Off-the-shelf tools spread that cost across many customers. Custom software can pay back when the process is how you win work and nothing on the market supports it without costly workarounds.
How do you calculate total cost of ownership for software?
Add setup, licences or build costs, maintenance, hosting, support, training and the manual work that remains after go-live, over three to five years. Leftover manual time is the cost most comparisons forget. Cost staff time at a realistic hourly rate and count the weeks people actually work, then compare each option's total rather than its first-year price.
When should a business integrate rather than buy new software?
Integrate when your current systems do their jobs well but people copy data between them. Connecting two trusted tools is usually cheaper and faster than moving everyone to a new platform. It also avoids retraining and data migration. Buy new software only when the existing tool can't support the redesigned process at all.
Who should make the software build vs buy decision?
The people who run the process, supported by someone who understands the technical options and has no product to sell. Include finance for the cost model and at least one daily user for fit. An independent assessment helps because vendors and development agencies each have an interest in one particular answer.
Business problem first. Technology second.
Weighing build, buy, integrate or automate for your next system? Book a first consultation with Severus