Off-the-shelf, configurable ERP or custom: what really changes
An off-the-shelf product is a closed system built for a common problem, such as invoicing, payroll or financial control. You adapt the company to it. It works very well when the problem is the same everywhere.
A configurable ERP is a large platform with inventory, purchasing, production and finance modules that you set up and, when needed, customize. It's the natural choice for anyone who wants a single registry of customers, products and suppliers. The decision between a custom ERP and a ready-made one is usually framed poorly: it's almost never one or the other. The most common and healthiest setup is an ERP handling what's standard and a purpose-built system handling what isn't.
Custom software is built from your process. You don't adapt the company to the system; the system is born the way your operation works. In exchange, you take on the cost of building and maintaining it.
When custom software for industrial companies makes sense
The first criterion is whether the process is a differentiator. If the way you inspect, plan a shutdown, put together a technical proposal or track a field service is part of the reason customers buy from you, forcing it into a generic mold removes exactly the part that's worth money.
The second is the symptom of permanent adaptation. When the team keeps side spreadsheets for everything the system doesn't do, or when every request to the product vendor turns into an expensive, slow customization, the system is already costing more than it seems.
The third is a combination of rules that exist only in your operation. In a system for industry, for example, it's common to have sequences of steps, approvals and records that change by type of equipment, by customer or by applicable standard. A generic product can handle that, but at the expense of improvised fields and people who have to remember the exceptions.
- The process is part of the company's value proposition.
- The team keeps parallel spreadsheets to make up for the current system.
- Customizations to the current product are expensive, slow or impossible.
- There are documents, reports or inspection certificates in a specific format, required by customers or by regulations.
- The system needs to talk to equipment, customer systems or databases that the product doesn't support.
When it doesn't make sense (and a ready-made product works better)
If your process is the same as any other company of your size, don't build. Finance, accounting, payroll, tax invoicing and corporate email are solved problems. Building them custom means spending money to end up with something worse and harder to maintain.
We also don't recommend custom software when the company still doesn't know how the process should work. Software doesn't organize a messy process; it freezes the mess. In that case, start by mapping the flow and testing with simple tools.
Finally, be careful with the urge to have the perfect system. Custom software requires someone at the company who makes decisions, validates screens and answers questions. Without that owner on the client side, the project runs late and comes out different from what was needed.
Integrations, reports and specific documents: where the math usually flips
In practice, the decision is rarely settled on the screens. It's settled at the edges: integrations, reports and documents. That's where a ready-made product shows its limits.
Integration is the first point. Ask whether the product has a documented API, what it lets you read and write, and whether there are usage limits. Integrating with a machine, a customer's system or one of your own legacy systems almost always becomes a separate project, and its price rarely makes it into the initial proposal.
Reports are the second. The standard dashboard covers 80% of the needs, and the other 20% become requests in the vendor's queue. If that 20% is what leadership looks at every week, it weighs more than the rest.
Documents are the third and the most underestimated. Inspection reports, certificates, service orders and technical reports in a specific format, sometimes defined by the end customer, are usually the point where a configurable system requires a parallel workaround in Word or Excel. A custom management system generates that document straight from the data, with no rework.
Total cost over the years: the comparison almost no one makes
Comparing the implementation price is easy and misleading. What decides it is the total cost over five to ten years, and it has more items than the proposal shows.
With an off-the-shelf product or ERP, add up licenses or subscriptions (which tend to grow with users and modules), implementation, customizations, integrations, training, version upgrades and the time your team spends working around limitations. With custom software, add up development, hosting, ongoing maintenance, evolution and the time your people spend validating the work.
There's no universal answer. In general, a ready-made product costs less at the start and custom software weighs more on the initial investment, but the outcome depends on the number of users, the volume of customizations and how much the system needs to change. If you want to understand the variables that drive the price, we wrote about it in how much a custom system or app costs.
An honest calculation also includes the cost of mistakes: hours lost to rework, duplicated information, decisions made with outdated data. It's hard to measure, but anyone who works in operations knows it exists.
Lock-in: the risk exists on both sides
When people talk about vendor dependency, they think of off-the-shelf products, and rightly so: if your data can only be exported in a closed format or the subscription price goes up, leaving is expensive. But custom software has lock-in too. If only the software house knows the code, you're tied to it.
The difference is that with custom software, this risk can be negotiated and reduced from the contract onward. Here is what we ask clients to require from any vendor, including us:
- Ownership of the source code and access to the repository.
- Minimum documentation of architecture, database and deployment.
- Data in open formats, with full export at any time.
- Common market technologies that other teams know how to maintain.
- An environment that third parties can reproduce, with no dependence on proprietary tools.
How to start with one module without betting the company
The most common mistake we see is trying to replace everything at once. The scope grows, the timeline doubles, and by the time the system goes live, the process has already changed. It's almost always better to pick one module and grow from there.
How to choose that first module: look for the point where the pain is greatest and the scope is small enough to be delivered in weeks, not years. It's usually a routine run in a spreadsheet, a document produced by hand or a step where information gets lost between departments. Define what a good result would be in simple terms, such as less rework or less time to issue a document, and only then begin.
Before writing any code, it's worth seeing the idea working. A free app prototype shows the screens and the flow so the team can validate the design before any larger investment, which avoids a good share of the most expensive rework.
If there's already an old system in the way, the option doesn't have to be throwing everything away. Legacy system modernization lets you replace it piece by piece while keeping operations running. That's the path we advocate in most cases, and we cover it in more detail in legacy system modernization: the complete guide.
For those who want the big picture of how to hire and what to expect from a software house, the guide on software development in São José dos Campos and the Vale do Paraíba covers the subject, using a region with a strong industrial base and technical services as the backdrop.
Questions to ask before you decide
Before requesting proposals, answer these questions with the team that uses the system day to day. If most of the answers are yes for the first ones, a ready-made product is probably enough. If they're yes for the last ones, it's worth talking about custom software.
- Is our process the same as that of companies similar to ours?
- How many spreadsheets and parallel controls exist today?
- Which documents and reports do we need to issue in our own format?
- Which systems or equipment will the new system have to integrate with?
- How much does it cost per year to work around the limitations of what we have?
- Who at the company will own the project and make decisions?
- If we switch vendors five years from now, what do we take with us?
From our experience
At SKAD, we keep custom systems running in areas where the client's process is specific: DEM, for equipment inspection; Inspefire, for fire safety inspection; SyncStop, for tracking industrial shutdowns; and Jecel, for inspection management and generating custom documents for the client's operation.
Inspefire is the example that teaches the most about starting in parts. Over 10 years, it received more than 10 new modules, added as the operation called for them, rather than from a giant scope defined on day one. Jecel, for its part, is still maintained by us, because for us delivering the system doesn't end the work.
Frequently asked questions
Is custom software always more expensive than an ERP?
On the initial investment, usually yes. Over five to ten years, the math can flip if the ERP requires many customizations, growing licenses and manual work to cover what it doesn't do.
Can I use an ERP and custom software together?
Yes, and it's a common combination. The ERP handles what's standard, such as finance and tax, and the custom system handles the process that sets the company apart, integrated with the ERP via API.
How long does it take to get the first module working?
It depends on the scope, the integrations and how available your team is to validate. A well-defined module is usually measured in weeks or a few months, not years.
What happens if the software house disappears or I want to switch?
If the contract guarantees ownership of the code, documentation and data export, another team can take over the system. That's why these clauses should be defined before you start.
Can service companies also benefit from a custom management system?
Yes, especially when the service involves inspections, reports, field visits or technical documents. In those cases, the workflow and the final document are usually the most specific part of the operation.
Conclusion
If you're still undecided, don't start by choosing a technology. Take a single process, the one that generates the most side spreadsheets, rework or hand-made documents, and describe on one page how it works today and how it should work. That page already tells you whether your case calls for an off-the-shelf product, an ERP or custom software. If it's custom, the cheapest next step is to see the idea working before hiring full development: ask for the free app prototype, validate the flow with the people who will use it, and only then decide the size of the investment.