Custom Software vs Off-the-Shelf: The 5-Year Cost

Every first call we take includes some version of the same question: is it cheaper to build our own, or keep paying monthly? The answer is usually not the one a software company is supposed to give. Here is the five-year arithmetic, using our own hosting and maintenance figures rather than a sales deck.

The Comparison Almost Everyone Gets Wrong

The sum people do in their head is a one-off price against a monthly one. RM 45,000 today, or RM 900 a month forever. Divide one by the other, get about four years, and conclude that building pays for itself before then.

That sum is wrong, because only one side of it is complete. The subscription price is genuinely all-in — the vendor hosts it, patches it, backs it up, and answers the phone. The build price is the cost of getting to launch day. It is not the cost of owning the thing.

Below is what the second column actually contains.

What a Custom Build Costs After Launch

Two recurring costs, and neither is optional.

Cloud hosting. We run client systems on AWS, billed pay-per-use rather than a flat plan. Across our own projects that averages around RM 600 a month, and it moves with the size of the system — a single internal tool used by six people sits well under that, a customer-facing platform taking real traffic sits above it. It also scales with your success, which is worth knowing in advance: the month your bookings double, so does part of that bill.

Maintenance. We charge roughly 15% of the build price a year, varying with complexity. That covers security patching, dependency and OS updates, the annual Apple and Google developer renewals if there is a mobile app, and the work of keeping something running that the outside world keeps changing underneath. It begins once the post-launch support included in your tier runs out — one month on Starter, three on Growth, six on Enterprise.

Changes are separate. A year in, someone always wants a new report, another user role, or a second payment method. We quote those by man-days, or absorb them inside a retainer if you are on one. Either way, budget for the fact that year two is not free. A build that never changes is usually a build nobody uses.

Five Years, Both Columns Filled In

Take a Growth-tier build at RM 45,000 — a real mid-range project, not a worst case. Against it, a subscription at RM 1,200 a month, which sits in the middle of the RM 200–2,000 range Malaysian SMEs typically pay for business software.

Building costs around RM 43,000 more over five years, before a single change request. That is the opposite of what the four-year payback sum suggested, and it is the number most vendors would rather you worked out after signing.

Move the inputs and the picture shifts, but not as much as you would hope. At the top of the subscription range — RM 2,000 a month, or RM 120,000 over five years — building finally edges ahead. At the bottom, RM 200 a month for a tool that does the job, subscribing wins by roughly RM 100,000 and it is not close.

The honest summary: custom software is rarely the cheaper option. It is the option you choose when the cheaper one does not actually do the job.

What the Subscription Price Also Hides

The table above is deliberately generous to the subscription column, because that is the honest way to make the comparison. But it is not the whole picture either, and three things belong on that side of the ledger.

The price is not fixed for five years. Nobody's is. Per-seat products get more expensive as you hire, which means the column grows exactly when the business is doing well. A tool at RM 1,200 a month for twelve staff is a different tool at twenty-five staff, and the five-year figure you calculated on today's headcount is the optimistic one.

You are renting the exit as well as the software. Ask, before you sign, what happens to your data if you leave — whether you can export customers, orders and history in a usable form, and what it costs. Plenty of Malaysian SMEs discover the answer during the migration, which is the worst possible time. Owning the system means owning the database; that is a real benefit, and it does not show up in a monthly price.

The product changes without asking you. Features get retired, interfaces get redesigned, and pricing tiers get restructured. Most of the time this is fine or even good. Occasionally the one workflow you depended on quietly stops working the way it did, and there is no version of the conversation where you get it back.

None of that reverses the arithmetic. It narrows the gap, and it explains why some businesses rationally pay more to own the thing.

The Cheaper Version of Building

There is a third column most people never consider, and it is usually the right one: build far less than you were planning to.

The RM 45,000 figure above is a platform. Most of the pain that sends a business looking for one is concentrated in a single process — the report that takes two days to assemble, the approval that lives in WhatsApp, the stock number nobody trusts. Solving that one thing is a Starter-tier build, and the five-year sum looks completely different.

Hosting is scaled with the size of the system rather than held flat, because that is how pay-per-use actually behaves — a small internal tool does not cost RM 600 a month, and a multi-branch platform costs more.

The top row is the interesting one. At around RM 50,000 over five years, a narrow build sits below a mid-priced subscription — and it does the one thing no subscription does, which is match your process exactly. It also leaves you the option of extending it later, once you have watched people use it, rather than guessing at the full platform on day one.

We learned this the expensive way. We once lost a deal by proposing the whole platform to a business that needed one process fixed, and what that taught us now shapes how we scope every project.

The Cost of Not Maintaining It

One shortcut deserves naming, because it is common and it is expensive.

The 15% is the first line people cut. The system works, nothing is obviously broken, and a year of skipped maintenance saves real money. What accumulates instead is invisible: dependencies drift out of support, the operating system moves on, a payment gateway changes its integration requirements, and mobile platforms deprecate the APIs the app was built against.

Nothing happens for a while. Then something does — usually an app store rejecting an update, or an integration that stops authenticating — and the work to catch up is not one year of maintenance. It is a re-platforming exercise priced closer to a new build, on somebody else's deadline.

If the 15% is not affordable alongside the build, that is a useful signal about the build, not about the maintenance. Either scope smaller, or subscribe.

When You Should Not Build — Three Real Examples

These are the enquiries we turn away, or redirect, most often. We would rather say so on a public page than in a meeting you had to drive to.

1. Accounting

The Malaysian market for accounting software is mature, competitive and cheap. The established products handle SST, statutory reporting and your auditor's expectations, and they have been doing it for years. Rebuilding that from scratch means paying five figures to reach a starting line someone else crossed a decade ago — and then owning the compliance updates forever. We have never recommended it, and we do not expect to.

2. F&B point-of-sale and stock

Same story, more so. There are a lot of POS and inventory products serving Malaysian restaurants, they are feature-rich, and they are inexpensive. If what you need is to ring up a sale, track stock and close the day, buy one this week.

The exception is real and worth naming: when you have several outlets and the problem is no longer the till but the gap between the tills — no single view of sales, stock moving between outlets on WhatsApp, month-end that never reconciles. That is a different problem from point-of-sale, and it is the one we build for.

3. An ordinary business website

If you need a site that explains who you are and lets people contact you, a website builder like Wix will do it, quickly and for very little. Paying a development company for that is paying for capability you are not using.

Two things change the answer. One is ordering: the moment the site has to take orders with your actual rules — set meals, time-limited items, pickup windows, deposits, tiered pricing — builders start fighting you. The other is strategy: sometimes an app exists for brand reasons, or to occupy a customer's home screen, and that is a legitimate commercial decision even when a website would technically suffice. Just make it deliberately, with the five-year number in front of you.

So What Does Justify a Build?

One thing, really: your process is genuinely yours, and it is worth money. Every off-the-shelf product encodes somebody's assumption about how the work is done. Where your business agrees with that assumption, subscribing is free money. Where it disagrees, you pay for the mismatch every day — in staff working around the software, in numbers that need reconciling by hand, in the report that has to be rebuilt in Excel each month.

In practice, four things push a Malaysian SME across the line:

• Pricing or approval rules a standard tool cannot express — tiered rates, deposits, multi-step sign-offs, rules that change by branch.
• Several locations that need one truthful view — where the pain is between the systems rather than inside any one of them.
• Integration with something you are not replacing — an accounting package or a POS that stays, with a layer built around it.
• A workflow that is the business — the thing you do differently from competitors, which is precisely what generic software flattens.

If none of those is true, the five-year table above is your answer. If two or more are, the mismatch is probably already costing you more than the gap in that table — which is the sum worth doing next.

Do the Sum for Your Own Business

Before you talk to anyone, including us, work out two numbers. First, what the current way of working costs in wasted hours — our ROI calculator turns that into ringgit in about two minutes. Second, the five-year total of the tool you would otherwise subscribe to, using the table above as the template.

If the wasted-hours figure is smaller than the build gap, subscribe. That is not us being modest, it is arithmetic.

If you do decide to build, go in knowing the shape of the price. We have written a fuller breakdown of what custom app development costs in Malaysia and what moves a quote from one band to the next, and you can see the kinds of systems we actually build before committing to a conversation.

And if you want the honest read on your own situation, tell us what you are trying to fix. If the answer is an off-the-shelf product, we will tell you which category to look at. We lose nothing by being right about that.