Build vs Buy: Should Your Business Get Custom Software or Off-the-Shelf?
A six-question build vs buy test for small businesses — when SaaS wins, when custom pays back, and how to decide in one sitting.
July 2026 · 9 min read
Buy when your process matches the product. Build when daily exceptions — duplicate entry, partial dispatch, account pricing — eat hours your team should spend on customers.
What build vs buy actually decides
Build vs buyChooses between subscribing to existing software and funding bespoke work that fits your workflows, integrations, and data. Most growing businesses end up with both.
You might buy a CRM and still need a custom enquiry router. You might keep your accounts package and still need a commerce bridge the middleware vendor never shipped. NuvonHub's custom software work usually extends tools you already pay for — because replacing working finance for a unified suite is rarely the fastest win.
The six-question build vs buy test
Answer honestly. Score one point for each "yes" to build. Three or more yes answers → plan a custom phase before buying another subscription.
| # | Question | If yes → lean |
|---|---|---|
| 1 | Do staff re-type the same data into two or more systems daily? | Build connector |
| 2 | Does more than 25% of revenue depend on workflow exceptions SaaS docs call "edge cases"? | Build |
| 3 | Would changing your process to fit software slow operations or frustrate key users? | Buy + bridge |
| 4 | Is there no mature SaaS category for your industry niche? | Build |
| 5 | Will integration middleware cost more than £6k/year in fees and admin time? | Build once |
| 6 | Does leadership need a dashboard no off-the-shelf report provides? | Build |
When buying wins
- Email marketing (Klaviyo, Mailchimp) — commodity capability.
- CRM for standard B2B pipeline (HubSpot, Pipedrive) — see our CRM comparison on day 7.
- Accounting when you are not on legacy desktop ERP.
- Ticketing and chat when workflows are industry-standard.
- Hosting, CDN, auth — never build these from scratch in 2026.
When building wins
- ERP ↔ commerce sync with partial dispatch and account pricing.
- Ops dashboards combining accounts exports, web orders, and stock flags.
- Catalogue reconciliation across 10,000+ SKUs with audit trails.
- Approval workflows unique to your compliance or franchise rules.
- Customer portals that mirror how trade buyers already order.
Worked example: buy the store, build the bridge
On a recent integration project, we kept the client's existing platforms and built reconciliation plus workflow automation — 13,602 Sage SKUs analysed, 7,295 inactive candidates flagged; removed or set inactive on web after client sign-off, duplicate entry reduced. Build cost focused on the logic middleware vendors hand-wave. That is the hybrid model most SMBs need but few agencies articulate.
Cost comparison over 24 months
| Approach | Typical spend | Risk |
|---|---|---|
| SaaS only + 3 integrations | £18k–£45k subscriptions | Workarounds multiply |
| Custom core + commodity SaaS | £25k–£55k build + £8k SaaS | Scope control needed |
| Full custom platform | £60k–£150k+ | Maintenance burden |
| Do nothing (manual) | £0 software + 500+ staff hours | Errors and delay |
Staff hours belong in the model. Ten hours/week of re-keying at £25/hour loaded cost is £13,000/year — before a single mis-invoice. Workflow automation ROI (day 6) walks through the payback math.
How to start without betting the company
- Map one painful workflow end-to-end — order to invoice is the usual winner.
- List every system touched and every manual step.
- Price buy options: licences + middleware + implementation.
- Get a fixed-scope quote for custom on that slice only.
- Pilot with one product line or customer segment.
- Measure hours saved and error rate for 30 days before expanding.
Series navigation
Previous: SaaS vs custom software. Next up: agency vs freelancer vs in-house for web projects.
Frequently asked questions
- Is custom software worth it for a small business?
- Yes — when manual work between systems costs more than £10k–£15k per year or when errors risk customer relationships. It is not worth it for commodity needs like email or standard CRM.
- What is cheaper: build or buy?
- Buy wins on year-one cash for standard workflows. Build wins on 24–36 month TCO when middleware, staff workarounds, and sync failures accumulate.
- How long does custom software take for an SMB?
- A focused connector or dashboard: 6–12 weeks. Broader operational platforms: 4–9 months. Phased delivery reduces risk.
- Can I build on top of SaaS instead of replacing it?
- That is the recommended path for most SMBs — keep mature SaaS for commodity features; build thin integration and rules layers around it.
- What does NuvonHub typically build vs buy for clients?
- We buy hosting, CRM, email, and commerce platforms where they fit; we build integration bridges, catalogue tools, ops dashboards, and automation that encodes client-specific rules.