The answer turns on one question: is the process you want to systematise your competitive advantage, or simply something that has to get done?
Whatever sets you apart from competitors may deserve custom work. Whatever everyone does the same way — invoicing, payroll, email — is bought off the shelf.
Not sure which category yours falls into?
Describe the process to us and we will tell you honestly whether you need development or just proper configuration.
The four-question test
Answer these before you look at any quote. They settle 80% of the decision.
| Question | If yes |
|---|---|
| Does a ready-made tool cover 70% of your needs? | Start there |
| Does your process differ substantially from your sector’s norm? | Custom starts to make sense |
| Will you change that process materially within a year? | Do not build yet |
| Do subscriptions already exceed €800 a month? | A comparison with development is worth running |
The third question is the most commonly skipped. There is no sense encoding a process that is still finding its shape.
When off-the-shelf wins
In most SME cases, and for specific reasons:
- It works from week one. Development takes months.
- The cost is predictable and spread across a subscription.
- Updates arrive on their own, including regulatory changes.
- Documentation and a community exist — you do not depend on one person.
- You can trial it before committing. Most platforms offer a trial period.
The most important point: a ready-made tool encodes the experience of thousands of businesses. If it does not do something the way you imagined, the first question worth asking is why.
The core categories almost always bought rather than built are covered in our guides to what ERP is and what CRM is.
When custom is justified
Four cases, and only these:
- The process is your product. A costing calculation nobody else performs that way, a routing system shaped by your specific geography.
- No ready-made tool comes close. You established that after examining at least five, not because the first one felt awkward.
- Volume makes subscriptions prohibitive. Platforms charging per user or per transaction become expensive at scale.
- Integration is the actual problem. You have five systems that do not talk to each other and you need the bridge — often the cheapest form of custom development there is.
The fourth is the most common in practice. See how it works in the case of connecting an ERP to an e-shop.
The real five-year cost
Comparing “€20,000 to build” against “€200 a month” is misleading. The honest comparison is this one:
| Component | Off-the-shelf | Custom build |
|---|---|---|
| Initial cost | €0 – €3,000 | €15,000 – €60,000 |
| Monthly subscription | €100 – €600 | €0 |
| Hosting and infrastructure | Included | €50 – €300 per month |
| Maintenance and fixes | Included | 15%–20% of build cost yearly |
| Configuration | €1,000 – €5,000 | Included |
| Training | €500 – €2,000 | €1,000 – €3,000 |
| Indicative five-year total | €12,000 – €45,000 | €30,000 – €100,000 |
The ranges are wide because they depend on user numbers and complexity. The conclusion holds regardless: custom rarely comes out cheaper before year three or four, and only if requirements stay stable.
And one factor that does not fit in a table: custom demands your involvement. Specifications, testing, decisions. That time has a cost.
The hybrid approach that usually wins
This is not a two-option dilemma. The most common right answer is the third one:
A ready-made platform as the base, plus custom development at the point where you differ.
| Layer | Choice |
|---|---|
| Accounting and invoicing | Off-the-shelf, always |
| Stock and orders | Off-the-shelf, configured |
| Your distinctive process | Custom |
| Integrations between them | Custom or an automation tool |
| Management reporting | A ready-made tool sitting on your data |
The advantage: you pay for development only on the 10% that genuinely differs, instead of rebuilding functions that have existed for twenty years.
Want to see how this applies to you?
Send us the list of tools you currently use and we will propose an architecture.
What vendor lock-in is, and how much it should worry you
It is the situation where changing provider costs so much that, in practice, it stops being an option.
The three questions to ask before signing:
- Can I export all my data? In what format, how often, at what cost.
- Is there an API? Without one, every future connection becomes disproportionately expensive.
- What happens if I stop the subscription? How long do I retain access to my history.
Watch out for one misconception: custom does not free you from lock-in. It simply moves it from the software vendor to the developer who wrote the code. That is why source code ownership and documentation must be written terms in the contract.
The five mistakes that cost most
- Building because “nothing ready-made suits us”. Usually it means two options were reviewed, not ten.
- Copying the existing process into code. Systematising is a chance to simplify, not to set the old way in stone.
- Specifications without priorities. When everything is essential, the project triples.
- No maintenance budget. Software without maintenance becomes a security problem within two years.
- Choosing a platform without testing it on real data. The demo always runs perfectly.
Frequently asked questions
How long does a custom system take to build?
For a narrow scope, 2–4 months. For a full management system, 6–12. Add 30% to the initial estimate.
Can I start off-the-shelf and move to custom later?
Yes, and it is the safer route — provided you secure data export from day one.
Who owns the code?
Whatever the contract says. Ask explicitly for full transfer of source code and documentation.
What is “no-code” and is it reliable?
Platforms that let you build applications without programming. They work well for internal workflows and prototypes, less so for high-volume critical systems.
How do I evaluate a development quote?
Ask for a phased breakdown with deliverables, not a single figure. And ask what is excluded.
The first step
For two weeks, log every repetitive task you do by hand. Next to each one, note how long it takes.
That list is your specification — and it usually proves the problem is solved by configuration and connection rather than new software. The same incremental logic is set out in the first steps of digital transformation.
Next step
Book a free 30-minute call with the team at The Dev Alley. We will look at your processes and tell you what is worth buying and what is worth building.