In this article (11 sections)
The discussion usually gets framed badly. "Custom" doesn't mean better and "off the shelf" doesn't mean cheap. They're two different ways of solving the same problem, with cost, risk and timeline structures that look nothing alike.
This is a concrete way to decide, with the numbers and the questions to ask yourself before committing to either path.
What you're buying in each case
When you buy an off-the-shelf product — an ERP, a CRM, one of the business systems sold by subscription — you're buying a process someone else already defined. The software embodies the way its maker believes work should be done, distilled from hundreds or thousands of customers. Adopting it means, to a large extent, adopting that way.
When you buy custom software, you're buying a translation of your current process into a system. Nobody is going to ask you to change how you work, because the system gets built around how you work.
That's the whole difference, and everything else follows from it.
The question that decides almost everything
Before comparing prices, answer this: is your way of working a competitive advantage, or just the way it ended up being done?
It's an uncomfortable question because almost everyone believes their process is special. In most cases it isn't: it's a standard process with quirks that piled up from old decisions, people who left, and limits that no longer exist.
- If your process is standard, an off-the-shelf product is almost always the better decision. You'll pay less, start sooner, and get improvements other people paid for.
- If your process genuinely is your advantage — the thing that makes you faster, cheaper or better than your competition — forcing it into someone else's mold destroys the reason people pick you.
How to tell: if a competitor copied your process exactly, would it help them? If the answer is "it wouldn't change a thing for them", your process isn't the advantage.
The numbers, including the part they hide
The real cost of off the shelf
The subscription is the visible part and usually the smaller one. A serious business management product for a small company lands between USD 30 and 150 per user per month. For twenty users, between USD 7,000 and 36,000 a year, every year.
What gets added and isn't on the pricing page:
- Implementation and configuration. It's rarely self-service. For an ERP, implementation usually costs one to three times the annual license.
- Data migration. Getting your information out of where it is and into the product's format.
- Modules charged separately. The price on the page is usually the base plan. Payroll, tax filing, integrations, advanced reports, read-only users: each with its own cost.
- The custom development on top of the off-the-shelf product. This is the irony. Most ERP implementations end with specific development on top, because there's always something the product doesn't do. And building on someone else's product is more expensive than building from scratch, because you're limited to what their platform allows.
- Certified consultants. The big products have a consultant ecosystem, and an hour of a certified consultant isn't cheaper than an hour of a developer.
The real cost of custom
- Initial build. For a small-business management system, USD 8,000 to 25,000 is where most of them land.
- Infrastructure. Between USD 20 and 200 a month depending on volume. Not per user: per system. That difference is key and I come back to it below.
- Maintenance. On the order of 15% to 20% a year of the build cost.
- Evolution. Each new feature gets quoted and paid for.
Where the curves cross
Off the shelf clearly wins the first two years. Custom has a bigger upfront outlay but a substantially lower recurring cost, and above all it doesn't grow with the number of users.
That last point is what flips a lot of decisions. A custom system for 20 users and for 120 users costs practically the same to run. A per-user license doesn't. If you're growing, do the math with the number of users you'll have in three years, not today's.
As an order of magnitude: in companies above 30 users with particular processes, the break-even point usually falls between year two and year four.
When off the shelf is clearly the answer
- Accounting, payroll and taxes. Never build these custom. They're regulated processes that change by law several times a year, and keeping that current is permanent work somebody else is already doing for you.
- You need to be running in weeks. A product gets implemented faster than one gets built.
- Your process is standard and you know it. There's no value in building your own version of something that already exists.
- You don't have anyone who can run a development project. Custom software needs someone on your side who decides. Without that, the project drifts.
- The upfront budget is the hard constraint. A subscription turns an investment into an operating expense, and sometimes that's the difference between doing it and not doing it.
When custom is clearly the answer
- Your operation doesn't fit any mold. It happens often in logistics, in production with particular processes, in services with their own pricing logic, and in any business with a twist the industry doesn't share.
- You already tried a product and it didn't work. This is the strongest signal of all. If you already paid for an implementation that ended in parallel spreadsheets, another similar product isn't going to fix it.
- You're paying for a lot of licenses on something you use at 20%. Very common: big products with dozens of modules, of which you use three.
- You need to integrate systems nobody integrates. Scales, sensors, plant hardware, your own old systems, the software of the supplier you work with.
- You're going to grow a lot in users. Per the arithmetic above.
The third option, which is usually the right one
In practice, a lot of companies don't choose: they combine. And the dividing line is clean.
Off the shelf for what's the same at every company. Accounting, payroll, tax filing, email, e-signature. These are regulated processes or commodities. There's no competitive advantage available there.
Custom for what's yours. The heart of your operation: how you plan, how you control, how you dispatch, how you quote.
The two connected by integration. This setup usually gives the best cost-to-result ratio, because it doesn't pay licenses for what it doesn't need and doesn't build what's already solved.
The condition for it to work is that the off-the-shelf product has a decent API. Ask that before buying the license, not after. A closed product condemns you to manual re-entry between systems, which is exactly the work you were trying to eliminate.
Four questions before deciding
1. How many users will I have in three years?
It changes the math more than any other variable.
2. What percentage of my operation would the product cover with no modifications?
If it's 80% or more, off the shelf has a good shot. If it's 50%, you're going to end up building on top, and that's the worst combination: you pay the license and you pay for development, with someone else's platform limiting you.
3. Can I get my data out if I want to leave?
Full export, in a usable format, without depending on the vendor. Ask it explicitly and ask to see it. It applies to custom too: the code and the database have to be yours.
4. What happens if the vendor changes the price or discontinues the product?
With licenses, that's a real risk you don't control. With your own software, the risk is that your development provider disappears, but the system is still yours and someone else can pick it up.
How not to get it wrong
The expensive mistake isn't choosing badly: it's choosing before defining the problem properly.
Before looking at products or asking for quotes, write down in two pages what your operation does today step by step, where it breaks, and what you want to be different. That document works for both things: to evaluate whether a product covers your case, and to get a build quoted.
Without it, every demo is going to look good and every quote is going to be impossible to compare.