Have something built yourself or rent a standard solution? A decision almost every SME makes once - often in the wrong place. Six points that help you make it in the right one.
Whoever needs software today has two fundamental ways: rent an existing solution - or have your own one built.
We recommend both, depending on the situation. Custom software isn't always the right answer, even though we build it ourselves. SaaS isn't always the worse choice, even if it feels more interchangeable. This page helps you make the honest decision.
SaaS means Software as a Service - you rent software that someone else built and runs. It runs in their data centre, you use it through the browser. You pay monthly per user. Classic examples: Lexware Office, DATEV Cloud, Microsoft 365, Salesforce, Pipedrive.
Custom software means: you have your own software built that fits your business exactly. You pay once for the build and then ongoing for maintenance and hosting. It runs on a server of your choice (or in the cloud) and, ideally, belongs to you.
In between there are mixed and configuration solutions, but the two poles help to sort out the choice.
SaaS is the faster, cheaper, less risky choice everywhere you need the same thing as a thousand other businesses do.
Examples: accounting (DATEV, Lexware) - the business is the same everywhere, nobody should build that themselves. Email and calendar (Microsoft 365 or Google Workspace) - standard, very mature, ridiculously cheap relative to the value. Project management, if your requirements aren't specific (Asana, Linear, Trello). Sales CRM for a classic B2B pipeline (Pipedrive, HubSpot). Payroll (Personio, DATEV).
Rule of thumb: if the market has a mature solution for your problem and you can honestly say you don't work substantially differently from others - take it. You save time, money and complexity.
Custom software becomes right when your business is genuinely different in one place - and that very place decides your success.
Examples: a trades business that manages its enquiries, appointments, jobs and times in an application built for its way of working - because every SaaS solution either does too much (confusing) or too little (topped up with Excel lists). A teacher tool for PDF marking that doesn't exist in the market like that. A royalty management for a company whose billing logic fits no standard software.
Rule of thumb: if your competitive advantage lies in a particular way of working that no SaaS maps - build your own software for exactly that place. For everything else, stay with SaaS.
The rarest answer is SaaS only. The second-rarest is custom only. The most common is: both at once, each in its right place.
Concretely: accounting in DATEV or Lexware. Email in Microsoft or Google. Payroll in Personio. And one bespoke, individually built piece for exactly the place where your business is different - your enquiry process, your production planning, your special CRM. The individual tools talk to each other through interfaces.
This hybrid architecture isn't only the cheapest, but also the most robust. If a SaaS provider goes bust or doubles its prices, that only affects one part of your system - not everything at once.
If you're unsure whether a sub-process should be SaaS or custom, this little checklist helps.
One: Is there a SaaS in the market that does exactly what we need - without us having to contort ourselves? If yes: SaaS.
Two: Would the SaaS force us into changes in our way of working that we don't want? If yes: check custom.
Three: Does the SaaS have an open interface through which we can exchange data with our other systems? If no: dangerous, lean towards custom.
Four: If the SaaS triples its prices or cuts features in five years - could we survive that? If no: custom.
Five: Is our data so sensitive that we never want it in the cloud of a US provider? If yes: custom on our own server.
We build custom software - but we tell clients very regularly: "You don't need this piece from us. Take SaaS X. What we build is the surrounding part." That isn't weakness, that's honesty. Whoever uses custom software even where SaaS would do wastes money - and makes themselves needlessly complex.
Our typical recommendation looks like this: a well-chosen foundation of three to five SaaS tools, plus a bespoke piece for your genuinely own way of working - often the layer between the SaaS tools, or the enquiry management, or a specific production logic. The mix fits 80 percent of SMEs better than any pure form.
Choosing the right SaaS tools is a discipline in itself. Some seem cheap and turn out expensive, because they add up quickly per user. Others seem powerful and are, in practice, too complex for a small team. We usually help our clients with this choice too - without commission from the providers, so that we can advise honestly.
A good test: when you trial a SaaS, don't let the managing director test it, but the employees who'll later work with it. What shines in the sales pitch can feel dull in daily use - and only those who click daily notice the difference.
The right answer is rarely either or. It's usually which part where.
A good software partner recommends SaaS where SaaS is enough - even if they build custom themselves. Whoever wants to build everything for you is optimising their business, not yours. At the first meeting, watch whether someone is willing to recommend standard solutions to you. That's the most honest quality signal.
What custom software costs is on What software really costs. How new software fits into existing SaaS tools is on Letting existing systems live on. Who owns the code in the end is on Who actually owns the code?.