Blog
Vendor lock-in with low-code: the honest answer and how to manage it
· 5 min read
You are considering having an application built on a low-code platform. The demo went well, the plan is clear, and then your finance director asks the question you had been expecting: "And if we want something else in five years? Won't we be stuck?"
The honest answer is: yes, to some extent. That applies to low-code, but also to an off-the-shelf package and to custom software. The question is not whether you become dependent, but on what, how expensive leaving will be, and what you can do in advance to limit that.
Where does the dependency actually lie?
Lock-in is not a single thing. With low-code there are at least four layers, and they do not carry equal weight.
The data. This is usually the most valuable part of your application: customers, orders, installations, certificates. Data in an ordinary database can be migrated. How easily depends on how the platform stores the data and what access you have.
The logic and screens. With low-code you draw processes and screens in a visual environment. That model is the result of your business knowledge, but it is written in the language of the platform. You cannot simply take that language with you to another system.
The knowledge in people. Whoever manages the platform needs to know it. For an SME with one or two people who understand the platform, that is a real risk, and the same goes for custom software in an ordinary programming language.
The contract. Term, notice period, what happens to your environment at the end, and whether you can take your data with you. This part is the easiest to influence, and it is the one most often forgotten.
What does OutSystems itself say?
OutSystems is firm on this point. Its evaluation guide states that the platform does not use a proprietary runtime engine or proprietary interpreter, and that it converts your application models into standard .NET applications that can run on an ordinary web server. There it even calls itself the only solution that truly offers no vendor lock-in.
That sounds reassuring, but read the details. The starting point is the so-called detach process, which OutSystems describes on a dedicated support page. It explains what happens if you really leave:
- It only works for the .NET variant of OutSystems 11, from version 11.18.1 onwards.
- It is all or nothing. You cannot detach a few applications and leave the rest running on the platform.
- It is a one-way street. As soon as you start maintaining the generated source code yourself, you cannot go back to the visual model.
- You lose the platform tooling: the visual development environment, automated deployment and monitoring. Changes you make to the code yourself are no longer supported by OutSystems.
Our honest reading: you do not need to fear that your application will stop overnight, and you own a working .NET application. But you are also buying a codebase that was generated and was meant to be maintained through the platform. Whoever takes it over becomes the owner of a maintenance job. "No lock-in" is a marketing line. "An exit that costs time and effort" is more realistic.
Also pay attention to the version. This description applies to OutSystems 11. OutSystems also has a newer platform, OutSystems Developer Cloud. We have not found an equivalent, public description of the exit route there. Ask your supplier for it in writing before you sign, and include the answer in your agreements.
Five measures that genuinely reduce the risk
We build on OutSystems ourselves, including the certification platform for MEA Beveiliging, and we work with these measures.
- Keep the knowledge of your process outside the platform. Record what the application does: which roles, which screens, which integrations, which rules. Start with a mockup that you keep yourself. If you ever have to rebuild, you will then have a specification instead of a puzzle.
- Make sure your data is always accessible. Our advice: for complex applications, keep the critical data in a place you manage yourself, and record how you get the rest of your data back when the contract ends. That is an architectural choice you make up front and can hardly make up for afterwards.
- Integrate through standard interfaces. Let other systems talk to your application through common APIs, not through platform-specific tricks. A later replacement then does not affect the rest of your landscape.
- Keep the application modular. A number of smaller parts with clear boundaries is easier to replace than one big tangle. That applies to any system, but with low-code it is tempting to skip because building goes so fast.
- Arrange the exit in your contract and test it. Ask what happens to your data and your environment at the end, how you get them and in what format. Do a trial export once a year and see whether you can actually use it.
And what about the law?
Regulation from Brussels helps too. The Data Act contains rules that make it easier for customers to switch between providers of data processing services, and it has applied since 12 September 2025. Whether your low-code platform falls under it depends on how it is offered. We are not lawyers, so we do not claim this for any specific platform. It is, however, a good reason to ask your supplier how they implement these rules, and to have that put in writing.
So is lock-in a reason not to do it?
Usually not. Compare it with the alternatives. An off-the-shelf package ties you to the logic of the package and to the supplier of the customisations. Custom software ties you to the developer who wrote it, and to code that only that person understands. Also read our decision guide for custom software, an off-the-shelf package or low-code and the piece on when OutSystems does and does not fit.
So the question is not "lock-in, yes or no", but what the dependency gives you and what it costs to get out. An application that is up and running in weeks instead of months, and that genuinely automates your process, can be well worth that dependency. On the condition that you have looked at the exit route in advance.
What does this mean for your business?
- Write down what you will depend on. Go through the four layers: data, logic, knowledge, contract. For each layer, indicate how bad it would be if the supplier disappeared tomorrow or changed its terms.
- Ask every supplier three questions, including us. How do I get my data out, in what format and within what time frame? What happens to my environment when the contract ends? Can I have that in writing?
- Keep the specification yourself. Mockups, roles, integrations and business rules belong to you, not to the platform.
- Schedule an exit test. A trial export once a year is a small investment that prevents a big surprise.
Would you like to know what this looks like in your situation, or do you have an existing system you seem unable to get away from? Get in touch with us. We are happy to take an honest look, even if the conclusion is that low-code is not the right choice.