Buy off-the-shelf for common, non-differentiating needs. Build custom when a process is core to how you compete, when off-the-shelf forces costly workarounds, or when licence and manual-effort costs outweigh a purpose-built system. The numbers say most businesses are already over-buying: organisations run 150–300 SaaS apps on average, roughly half of paid licences sit idle, and 15% of apps are never used at all — the real question is where a tailored system gives you an advantage and where generic is good enough.
Most New Zealand businesses accumulate software the way a garage accumulates tools — one purchase at a time, until you have a dozen subscriptions and your actual problem still isn’t solved. The data confirms the pattern. Zylo’s 2026 SaaS Management Index puts the average organisation at 305 SaaS applications (small companies at around 152 in the 2025 edition), with licence utilisation at just 54% — meaning nearly half of what’s paid for isn’t used. Vertice’s 2025 shelfware benchmark, cited by Breeze, found 15% of applications entirely unused and 51% under-utilised. And 79% of IT leaders faced price increases at renewal last year.
So how do you decide when to keep buying off-the-shelf and when to build something custom? Here’s the framework we use with clients.
Off-the-shelf is the right call when…
- The need is common and not unique to you — accounting (Xero, MYOB), email, payroll, document storage, video calls.
- A well-supported product fits your process closely without heavy workarounds.
- You’re happy to adapt how you work to how the tool works.
- The vendor is stable, the data is exportable, and the price is predictable — check all three before you sign.
For these, buying is faster and cheaper. There’s no sense building what already exists. The trap is not buying off-the-shelf; it is buying too much of it and stitching it together by hand.
Custom software is the right call when…
- The process is core to how you compete — the thing you do differently and better. A trades firm’s job-management flow, a wholesaler’s pricing logic, a clinic’s triage rules.
- Off-the-shelf tools force awkward workarounds, or you’re stitching several together with manual effort in between. (If someone re-types data between two systems every day, you are already paying for custom software — in wages.)
- Per-seat licence costs and lost time add up to more than a purpose-built system would over three to five years.
- You need systems to talk to each other in ways generic tools won’t, or you need to embed AI agents into your own workflow rather than a vendor’s.
- Data residency, Privacy Act 2020 obligations or a customer’s security requirements rule out the available SaaS options.
When your business has outgrown generic tools, custom software shaped around how you actually work removes the friction — and you own it, with no per-seat licence trap. In Datacom’s 2026 index, 26% of NZ organisations now run custom-built or in-house AI, up sharply, for exactly this reason: the differentiating process is the one worth owning.
The hidden third option: integration
Sometimes you don’t need to replace your tools — you need them to work together. Enterprise integration and purpose-built CRM & ERP solutions can connect and extend what you already have, which is often cheaper than a full rebuild. Most of the “we need custom software” conversations we have end up as “keep Xero and your CRM, build the 20% in the middle that nothing off-the-shelf does, and integrate the lot.”
A five-year cost comparison
Licence fees look small next to a build quote until you add the years and the hidden labour. A worked example for a 25-person business:
- Off-the-shelf, with workarounds: three overlapping tools at NZ$60–120 per user per month ≈ NZ$45,000–90,000 over five years, plus the manual glue — say one person spending a third of their week re-keying and reconciling, ≈ NZ$25,000 a year, NZ$125,000 over five years. Total: NZ$170,000–215,000, and you still own nothing.
- Custom system for the core process, integrated with Xero and the CRM: a fixed-price build plus modest hosting and support. Depending on scope, the five-year total is typically comparable or lower — and the manual glue disappears in year one.
The numbers will differ for your business; the method is the point. Count the licences, count the hours between systems, and compare over five years — not one.
The test isn’t “custom or off-the-shelf?” — it’s “where does a tailored system actually give us an advantage, and where is generic good enough?”
A quick decision checklist
- Is this process something customers choose us for? → Build (or build the differentiating part).
- Does a mature product do 90% of it without workarounds? → Buy.
- Do we already own good tools that just don’t talk to each other? → Integrate.
- Are we paying for seats nobody uses? → Audit first; half of licences typically are.
- Would the vendor’s price increase or shutdown hurt us badly? → factor that into Build.
Deciding with confidence
The wrong build is expensive; so is the wrong workaround, paid for a little at a time. An independent look at your options — build, buy, or integrate — is exactly what our technology consulting is for. We scope it fixed-price, so you can decide with clear numbers rather than a sales pitch. If the answer is “build”, custom software development is delivered the same way: scoped, fixed price, and yours.
Sources
- Zylo, 2026 SaaS Management Index — application counts, licence utilisation, renewal price increases
- Vertice, SaaS Wastage and Shelfware Benchmark 2025, and Zylo 2025 small-company figures — via Breeze
- Datacom, 2026 State of AI Index — custom-built AI adoption in NZ organisations
Worked example uses indicative New Zealand costs; your figures will vary. Last updated 15 September 2026. Written by Ashok Poshamalla, Founder, Avtrix Software Solutions, Taupō.