Total cost, dependence, data and integrations: four criteria and three options for deciding whether to buy, add your own layer or build.
“Do we build a custom application or buy a subscription?” The question comes up at almost every company that has outgrown spreadsheets. The right answer is neither “custom” nor “SaaS”: it depends on how specific your process is, how much time it costs you today and how much you want to depend on someone else. This article offers a decision framework, not a recipe. We have no hidden interest in always recommending bespoke software: often a subscription solves the problem better.
Start with the right question
Not “what is cheaper?”, but: does this process set me apart from competitors, or is it a common necessity? Invoicing, email, basic bookkeeping and calendars are common necessities, and the market has mature products. Your own way of pricing a quote, scheduling teams or serving customers may be exactly what makes you different.
The starting rule: for what is common, buy; for what sets you apart, consider building, but only if you have a reason to.
The four criteria
1. Total cost, not today’s quoted price
A subscription looks cheap at first and gets pricier through added users, extra modules and volume limits. Custom software looks expensive at first, but costs after launch through maintenance, hosting and improvements. Compare over a horizon of several years, with every line:
- subscription: price per user, modules, integrations paid separately, training, time lost working around limits;
- custom: development, hosting, maintenance, security updates, your time for specifications and testing.
If a supplier tells you a bespoke application “costs nothing after delivery”, that is a claim to check: software needs upkeep, which we cover separately in this series.
2. Dependence
With SaaS the supplier decides price, features, limits and the pace of change. It can raise prices, remove a feature you rely on or disappear. With custom software you depend on the team that knows it: if you part ways with the developer and do not hold the code, the documentation and server access, you are just as stuck.
Useful questions for both options: can I export my data completely and in an open format? Who owns the code? What happens if the supplier withdraws?
For cloud services, the EU Data Act, applicable from 12 September 2025, introduced rules on switching between data processing service providers, including export and open interfaces. It does not replace checking your own contract; ask a lawyer how it applies in your case.
3. Data
Where does the company’s data live and who can see it? SaaS means data at the supplier, with a processing contract, jurisdictions and retention rules you must read. Custom software lets you choose where the data sits (your own server, an EU cloud) and keep it as long and in the way you need, but it also puts the responsibility for security and backups on you.
If you process personal data, these choices are also GDPR compliance matters, not just preference.
4. Integrations
This is often where the argument is settled. A standard product integrates well with what many others already have and poorly with what only you have (special equipment, a rarer ERP, an internal process). If most of the value comes from linking three systems together, you may not need a new application but a connecting layer between what you already have.
A decision in three variants
- Buy and use as is. For common processes. You adapt the process to the product.
- Buy and add your own layer. You stay on a mature product, but connect automations, reports or small applications to it through its API. Often the variant with the best balance of effort and risk.
- Build. When the process is truly specific, existing products force big workarounds and you have a clear reason the investment is worth it.
A hypothetical example: an installations company uses a common accounting package (variant 1), a bought CRM to which it has linked a quote request form and its own pricing calculation (variant 2), and might build only that pricing calculation if its rules become too complicated for a spreadsheet.
Signs you should consider custom
- for years you have built workarounds (spreadsheets next to the product) for something the product does not do;
- you regularly lose person-hours copying data between systems;
- your business rules are too specific for the available settings;
- you need control over the data or where it sits;
- subscription costs rise with user numbers faster than the value you get.
Signs you should stay on SaaS
- the process is common and the product does it well;
- you have no one to handle maintenance or to support you contractually;
- the need is still unclear and you want to learn on a ready-made product;
- the gap between how you work and how the product works is small.
Before signing anything
Ask for a list of what you concretely receive (code, documentation, access, the right to change supplier), a cost estimate over several years, an exit plan and a small first version, so you can test the assumption before the big investment.
If you are at a crossroads like this, ME Digital provides custom software and maintenance, and we say openly when building is not worth it; write to us on the contact page.