Business Apps Without Coding: When No-Code Is Enough and When It Lets You Down
“Business apps without coding” is one of the most common requests companies bring to us — and it's a perfectly legitimate one. No-code tools can do a surprising amount today, and in many cases paying for development would be silly. At the same time, they have a fairly sharp limit beyond which the savings turn into a trap.
This article is an attempt at a fair map: what no-code can do, where it stops being enough, what it really costs and how to leave the platform once you outgrow it.
What you can build without a developer today
| Tool | What it's good for | Typical price |
|---|---|---|
| Airtable / Notion | records, simple databases, shared overviews | CZK 250–600 / user / month |
| Microsoft Power Apps | business apps on top of Microsoft 365 data | from ~CZK 500 / user / month |
| Glide / Softr | simple mobile and web apps on top of a spreadsheet | CZK 250–800 / user / month |
| Make / Zapier / Power Automate | automation and connections between systems | hundreds to thousands of CZK / month |
Realistically, these tools cover most of what lives in Excel in companies: asset and order records, approving purchase orders and holidays, attendance, a simple CRM overview, collecting data from the field via a form.
If you're at the stage where the main pain is “we have it in five spreadsheets and nobody knows which one is current”, start with a no-code tool. You'll build it in weeks instead of months, pay a fraction of the price and, above all, quickly find out what the process should really look like. That alone is enormously valuable.
Where no-code hits the wall
The limits don't arrive gradually — they arrive all at once. Most companies run into at least one of these five:
1. Price per user. Licences are paid per head. Up to ten people it's cheap; with fifty you're paying tens of thousands a month for a tool that still doesn't do exactly what you need. It's the same maths as with an off-the-shelf CRM — the cost grows linearly with the team, while the value doesn't.
2. Logic the platform can't handle. As long as the rules are along the lines of “if the amount is over CZK 50,000, send it to the manager”, everything is fine. Once you need a calculation based on your own algorithm, document versioning or a non-trivial state machine, you start gluing logic together across three tools and four webhooks.
3. Performance. No-code platforms are built for thousands of records, not hundreds of thousands. A sluggish data view that your team opens fifty times a day is a hidden cost you'll never see on an invoice.
4. The customer-facing part. An internal tool can look like a spreadsheet. An app you give to your customers can't — you need your own look, your own domain, sign-in under your brand and a presence in the App Store or Google Play. That's where no-code runs out fastest.
5. An integration that doesn't exist. Connectors cover the big global systems. A connection to Czech accounting software, a production line or an older internal database usually isn't among them — and working around it with exports and imports is exactly the manual retyping you wanted to get rid of.
The hidden cost nobody talks about: the in-house admin
No-code doesn't mean “no work”. It means “no developer”. Every company has someone handy who builds the system — and from that moment on, they're its unofficial administrator. A year later, that person spends several hours a week fixing automations and explaining why something didn't sync.
Two problems: first, it isn't their job, and nobody counts it as a cost of the system. Second, when they leave, the only documentation that ever existed leaves with them — in their head. We've seen companies rebuild a working system from scratch because of this.
Vendor lock-in: the data yes, the logic no
This is the most common unpleasant surprise. The platform will let you export your data — it's in the contract and it usually works. But apps aren't just data. Automations, forms, permissions, rules, links between tables and views exist only inside the platform and don't export anywhere.
When you leave after two years, you're not moving the app. You're rebuilding it and taking only the tables with you. The cost of leaving therefore grows in direct proportion to how many processes you've moved there in the meantime.
It works the other way round too: that's why it's good to know in advance which processes you'll put into no-code and which you won't.
How to decide: one question
Of all the criteria, one has proven itself to us over the long run:
Is this process something you do differently from your competitors — and does it earn you money?
If not — holiday tracking, invoice approval, a shared contact list — it's a standard process and no-code is the right answer. Programming a standard process is a waste.
If yes — your particular way of planning jobs, pricing, working with clients — you'll hit the limits of the template sooner or later. A template is by definition built for the average company, and your advantage lies in not being average. That's where a custom system pays off.
The way out: don't rewrite everything at once
The most expensive mistake when moving off no-code is deciding to “rebuild the whole thing properly”. It takes months, costs the most, and in the meantime your team works in the old system that nobody is developing any more.
What works is the opposite:
- Find the process that strains the platform the most. You'll recognise it by having the most workarounds and the most complaints.
- Build a custom app just for that process, connected to the data that stays in the no-code tool.
- Let the rest keep running and migrate gradually, based on what else hits the wall.
And above all: the no-code prototype you've been using for a year isn't wasted work. It's the best technical specification you could have — unlike a document full of wishes, it shows how the process really runs, where people get stuck and what nobody uses. We describe how to turn it into a specification in our article on how to write a technical specification.
What a custom app costs instead of no-code
For comparison with per-user licences: a simpler business web app starts with us at CZK 1,990 a month, a system with roles, workflows and an accounting integration at CZK 4,900 a month, and a complex enterprise system at CZK 9,900 a month. The subscription is paid for the system, not per head — if your team grows from ten to fifty people, the price stays the same. You'll find the detailed price bands in How much does a web app cost.
The break-even point usually falls between 15 and 20 users. Below it, no-code is cheaper and more sensible; above it, a custom system starts to come out ahead — and that's before you count the in-house admin's time.
Summary
| No-code | Custom app | |
|---|---|---|
| Time to launch | days to weeks | weeks to months |
| Cost with 5 users | low | higher |
| Cost with 50 users | high (grows with the team) | same as with 5 |
| Custom logic and look | limited | no limits |
| Customer-facing apps | no | yes |
| Moving to another solution | data yes, logic no | the code and data are yours |
The best strategy for most companies isn't choosing between the two, but getting the order right: start with no-code, get to know the process, and build a custom system only where it pays for itself.
Not sure which side of the line you're on? Get an indicative price in the configurator, or book a consultation — we'll go through your process and tell you straight whether no-code is enough or a custom app already makes sense.
