WhatsApp us

Blog6 min read29 Oct 2026

Build vs buy vs custom software: a decision tree for SMEs

Buy when a standard tool fits, configure when it almost fits, build custom when the workflow is your edge or tools do not connect. A five-question decision tree.

Build vs buy vs custom software: a decision tree for SMEs

By KYN AI Advisory Team — AI implementation specialists, Singapore

Buy a packaged tool when your process matches how the tool works. Configure or extend it when it almost matches. Build custom software when the workflow is how you compete, when several tools must be stitched together by hand, or when per-seat pricing grows faster than the value. Most SMEs end up with a mix: bought tools for commodity jobs and one custom layer for the work that is theirs.

This article gives you a five-question decision tree, the costs to compare, and the cases where each option fails. It assumes a business with 10 to 200 staff.

What are the three options?

Buy. A packaged product with a subscription. Examples: an accounting package, a CRM, an HR system. You adopt its way of working.

Configure or extend. A packaged product plus its own add-ons, automations and integrations. You bend the tool within its limits.

Build custom. Software built around your workflow, running in your cloud account. You design the way of working, and the code and configuration pass to you.

"Build" in this article means commissioning a build from a specialist, not hiring an in-house development team. The in-house route is a fourth option and a larger commitment.

The decision tree

Answer in order. Stop at the first question that gives you a clear answer.

Question 1: Is this a commodity job?

Payroll, general ledger, email, file storage and e-signature are done the same way in most businesses. A standard tool is cheaper and safer than anything you build.

  • Yes: buy. Stop.
  • No: go to question 2.

Question 2: Does a tool fit nearly all of the workflow without workarounds?

Aim for a fit where the gaps are small enough to live with. List the steps of your workflow. Mark each as supported, workaround or unsupported.

  • Mostly supported: buy, then configure.
  • Many workarounds or unsupported steps: go to question 3.

Question 3: Is the workflow how you win or how you serve customers differently?

If a customer would notice and value the difference, it is part of your edge. Standard tools make you look like every competitor.

  • Yes: lean toward custom.
  • No: go to question 4.

Question 4: How many tools does the work cross?

If one task touches email, a spreadsheet, an accounting package and a messaging app, and a person moves data between them by hand, the cost is in the gaps.

  • Three or more tools with manual moves: custom integration layer or custom system.
  • One or two tools: configure or extend.

We cover this pattern in why small businesses struggle with fragmented software.

Question 5: How does the cost look over three years?

Compute both sides, not just year one.

  • Buy or configure: subscription × seats × 36 months, plus add-ons, plus the labour of the workarounds.
  • Custom: build fee, plus a monthly retainer for at least twelve months and any ongoing changes, plus hosting.

Per-seat pricing grows with headcount. A custom system is priced on scope, not on seats. If your team will double, the comparison moves toward custom.

How do the options compare?

| Factor | Buy | Configure or extend | Custom | |---|---|---|---| | Time to first use | Days | Weeks | Under 90 days, after a 14-day review (KYN) | | Upfront cost | Low | Low to medium | Highest | | Fit to your workflow | Tool's way | Mostly | Yours | | Control of data and hosting | Vendor's | Vendor's | Your cloud account | | Change speed | Vendor roadmap | Within limits | As needed, via retainer | | Exit | Export, retrain | Same | Revoke the builder, keep running | | Main risk | Misfit, price rises | Fragile workarounds | Scope creep, bad build partner |

The "exit" row matters. With a bought tool, you leave by exporting data and retraining staff. With a custom system, the exit depends on who owns the cloud account and the code. Ask before signing. In our builds every system runs in an account in the client's name.

When does buying fail?

  • The workaround pile grows. Each workaround is a manual step that someone must remember.
  • Data is trapped. You cannot get your records out in a usable form.
  • Pricing moves with your growth. Seats, contacts or volume push the bill up faster than revenue.
  • Tools do not talk. You pay for three subscriptions and a person to copy between them.

See how this plays out for CRM in custom CRM vs HubSpot.

When does custom fail?

Be honest about this side too.

  • The scope is vague. "Build us something like Salesforce" is not a scope.
  • Nobody owns it after launch. Systems need changes. A retainer or an internal owner must exist.
  • A standard tool already fits. Building a replacement for a tool that works is waste.
  • The builder keeps the keys. If the account, code or configuration belongs to the builder, you have rented a system with extra steps.

What does a mixed approach look like?

Most SMEs should land here.

  1. Keep bought tools for commodity jobs: accounting, payroll, email.
  2. Build one custom layer for the work that is yours, such as enquiry handling, quoting or operations reporting.
  3. Connect the two so data moves without retyping.

For example, a manufacturer we worked with reduced eleven systems to three on one AWS account. A finance software startup went from nothing to a production receipts, claims and GST product in six weeks on Google Cloud. Those are different shapes of the same decision: one consolidated, one built from scratch because no tool did the job.

What should you do this week?

  1. List the five workflows that take the most staff time.
  2. Mark each as commodity, almost fits, or yours.
  3. For the ones marked "yours" or "almost fits with many workarounds", count the tools involved and the manual steps.
  4. Pick one. Put a number on its monthly cost in hours.

That number tells you how much a build can be worth.

Where KYN fits

KYN builds custom systems for SMEs in Singapore, Hong Kong, Indonesia and Malaysia, and we will say when a bought tool is the better answer. The first 14 days are a review: an inventory of your stack and interviews with your team. A full build ships in under 90 days. Pricing is a one-off build fee plus a monthly retainer of at least twelve months, quoted after the review.

Every system runs in a cloud account in your name. Rights in the custom code and configuration pass on full payment, and you can revoke KYN's role while everything keeps running. See infrastructure and pricing.

Frequently asked questions

Is custom software only for large companies?

No. The cost is driven by scope, not company size. A single well-defined workflow can justify a custom build at 10 staff if the manual cost is high enough.

Can I start with a bought tool and move to custom later?

Yes, and often you should. Keep your data exportable and your processes documented. A later custom layer can read from the bought tool through its API.

What if the builder disappears?

That is why ownership terms matter. If the cloud account is in your name and the code and configuration are yours, another engineer can maintain the system.

Checked 30 October 2026.

Curious whether Marketing agents fits your business? Talk to KYN on WhatsApp — no forms, just a conversation.

Start Building →

Written 29 Oct 2026 · Checked 29 Oct 2026