Build vs buy: custom software or off-the-shelf SaaS?

Should you build or buy software?

Buy, unless you have a specific reason to build. Off-the-shelf software is cheaper to start, faster to roll out, and maintained by someone else. The reasons to build are narrower but real: the process is how you win, nothing on the market fits without contortions, you need deep integration across systems, or subscription fees at your scale cost more than ownership.

A software developer telling you to buy might sound odd. It’s the honest starting point. For accounting, payroll, email, HR, standard CRM and project management, well-established products exist and a custom build almost never makes sense.

What’s the difference between custom software and off-the-shelf software?

Buying rents a product built for many customers; building creates an asset built for you. Off-the-shelf software (sometimes called COTS, commercial off-the-shelf) includes SaaS subscriptions and packaged products; custom software, also called bespoke software, is written for one organisation. Each wins on different things.

	Off-the-shelf SaaS	Custom software
Upfront cost	Low: setup and configuration	High: discovery, design and build
Ongoing cost	Subscription, usually per user, rising with headcount and price changes	Hosting, maintenance and enhancements
Time to value	Days to weeks	Months
Fit to your process	You adapt to the product	Built around your process
Differentiation	None: competitors can buy the same product	Can be a genuine advantage
Integration	Whatever the vendor’s APIs and marketplace support	Designed for your systems
Roadmap control	Vendor decides	You decide
Data location	Vendor’s choice of hosting, sometimes selectable	Your choice, including Australian regions
Maintenance burden	Vendor’s	Yours, or your developer’s
Exit	Depends on the vendor’s export tools	You own the code and data, if the contract says so

How do the five-year costs compare?

Compare total cost of ownership over at least five years, because year one flatters SaaS and hides the build’s long tail. The two examples below use illustrative figures to show the method. They’re not quotes; replace them with your real numbers.

Assumptions for both examples (AUD, ex GST): SaaS at $60 per user per month plus $20k for implementation, with a 5% price rise each year. Custom build at $250k, with maintenance budgeted at 18% of build cost per year ($45k) and hosting at $6k per year.

Example A: 80 users

	Year 1	Years 2 to 5	Five-year total
SaaS subscription	$57,600	$260,700	$318,300
SaaS implementation	$20,000		$20,000
SaaS total			about $338k
Custom build	$250,000		$250,000
Custom maintenance and hosting	$51,000	$204,000	$255,000
Custom total			about $505k

Subscription arithmetic: 80 users × $60 × 12 = $57,600 in year one, rising 5% a year to about $70,000 in year five. At this scale, buying wins clearly, and it would take a strong non-cost reason to justify building.

Example B: 400 users

Same assumptions, five times the users. Year-one subscription is 400 × $60 × 12 = $288,000; over five years with 5% rises it’s about $1.59m, plus implementation. The custom total is unchanged at about $505k, because hosting a few hundred more users rarely changes costs much for a business application.

At this scale, owning the software costs roughly a third of renting it over five years, and the gap widens every year after. This is the pattern behind most build decisions for core operational systems in mid-sized and large organisations.

Two caveats keep this honest. Custom software also needs change requests and new features, which the maintenance budget may not cover. And SaaS often includes things you’d otherwise have to build, such as mobile apps, reporting and compliance certifications. Count those on both sides. For typical build ranges, see how much custom software costs in Australia and software maintenance costs.

How do you decide? A weighted decision matrix

Score both options against the factors that matter to you, weighted by importance. The numbers force a conversation that “it depends” doesn’t.

Score each option 1 to 5 on each factor, multiply by the weight, and total the columns. Here’s a filled-in example for a logistics company deciding on a dispatch and job management system:

Factor	Weight	SaaS score	SaaS weighted	Custom score	Custom weighted
Fit to how we actually work	5	2	10	5	25
Five-year cost at our scale	4	3	12	4	16
Time to go live	3	5	15	2	6
Integration with our systems	4	2	8	5	20
Is this process a competitive advantage?	4	1	4	5	20
Internal capacity to own software	3	5	15	2	6
Data location and control	2	3	6	5	10
Total			70		103

In this example custom wins, driven by fit, integration and differentiation. Change the weights for a business whose dispatch process is standard and whose team has no appetite for owning software, and SaaS comes out ahead. The value of the exercise is agreeing the weights before anyone argues about the answer.

What about lock-in?

Both options lock you in; they just do it differently. Plan your exit before you sign either way.

SaaS lock-in comes from your data and your processes living inside someone else’s product. Check:

• Can you export all your data, including history, attachments and audit logs, in a documented format?
• What happens to prices at renewal, and is there a cap on increases?
• What happens if the vendor is acquired, changes direction or retires the product?
• How much of your process now depends on features only this vendor has?

Custom software lock-in comes from dependence on whoever built it. Protect yourself by:

• Owning the IP through a written assignment. In Australia, copyright in code belongs to the developer unless it’s assigned to you.
• Keeping the code in your own repository from day one.
• Using mainstream languages and frameworks that other teams can pick up.
• Requiring documentation and automated tests as deliverables, not extras.

What about the middle ground?

The best answer is often buy the core and build around it. Most organisations don’t face a pure choice.

Common hybrid patterns:

1. Integration layer. Buy standard systems (accounting, CRM, HR) and build the integrations that make them work as one, removing double entry.
2. Custom front end on bought back ends. A tailored portal for customers or field staff that reads and writes to off-the-shelf systems behind the scenes.
3. Build the differentiator, buy the rest. A custom pricing engine, scheduling optimiser or client portal sitting beside commodity software.
4. Extend with AI. Add an AI assistant or document processing to existing systems rather than replacing them. See the cost of adding AI to an existing app.

What should Australian organisations check?

Check where the SaaS vendor keeps personal information, what currency you’re paying in, and when support is available.

• Data location. Many SaaS products are hosted overseas. Disclosing personal information to an overseas vendor engages APP 8, and you can remain accountable for how that vendor handles it. Ask for the hosting region and the list of sub-processors.
• Security obligations. APP 11 requires reasonable steps to protect personal information, which includes choosing vendors with appropriate security. If a vendor has a breach involving your customers’ data, you may still have obligations under the Notifiable Data Breaches scheme.
• Currency exposure. Subscriptions priced in USD move with the exchange rate, which can add or remove a noticeable percentage from your annual cost with no change in usage.
• Support hours. A vendor supporting in US hours may answer your urgent ticket overnight. Check support hours in AEST and response commitments.
• Sector rules. Health, financial services and government buyers may face onshore hosting or specific security requirements that narrow the SaaS field.

Custom software lets you choose an Australian cloud region and your own support arrangement, which can simplify these questions. It doesn’t remove them: you’re then responsible for the security of the system. If you’re building something new, the R&D Tax Incentive may be relevant in some circumstances; speak to a registered R&D tax agent before assuming it applies.

Signs you’ve outgrown your off-the-shelf software

You’ve probably outgrown a product when workarounds cost more than the subscription. Common signals:

• Staff maintain spreadsheets alongside the system because it can’t handle a core step.
• You pay for several overlapping tools and re-key data between them.
• Per-user fees have become one of your largest software costs.
• You’ve asked the vendor for the same feature for years.
• Reporting requires exporting data and rebuilding it elsewhere.

How All Webbed Labs approaches this

We begin build-or-buy conversations by looking for a product that already does the job, and if one does, we’ll say so. Where custom software makes sense, we run a paid discovery that includes a five-year cost comparison against the best off-the-shelf alternatives, then quote a fixed price for the build. Your code lives in your repository from day one and the IP transfers to you on completion, so you’re never locked to us. See our enterprise software service.