No-code tools promise everything. Build apps without developers. Launch in days not months.

Partially true. The whole truth is messier.

What I Built

Project 1: Internal operations dashboard Built with Retool. Shows key metrics, automates reports. Status: Still running. Works great.

Project 2: Customer-facing booking system Built with Bubble. Appointment scheduling, payments, notifications. Status: Migrated to custom code after 18 months.

Project 3: Marketplace MVP Built with Glide and Airtable. Simple listing and matching. Status: Migrated to custom code after 8 months.

Different outcomes. Here’s why.

Where No-Code Works

Internal tools

Tools your team uses. Dashboards, admin panels, simple workflows.

Lower quality bar. You control the users. Bugs are annoying, not business-ending.

Retool, Notion, and Airtable handle internal tools beautifully.

MVPs for validation

Proving demand before investing in development.

No-code lets you test ideas in days. If the idea is wrong, you’ve wasted days not months.

When we validated our marketplace, Glide got us to our first 50 users. That was enough to know the concept worked.

Workflows and automation

Connecting systems. Moving data. Triggering actions.

Zapier and Make are legitimately excellent. Better than custom code for most integrations.

Simple CRUD apps

Create, read, update, delete. The basics.

If your app is mostly forms and lists, no-code handles it fine.

Where No-Code Breaks

Complex logic

When business rules get complicated, no-code gets ugly.

Our booking system had 40+ conditional rules for scheduling. In Bubble, this became unmaintainable spaghetti.

Custom code handles complexity cleanly. No-code doesn’t.

Performance at scale

No-code platforms optimise for ease, not speed.

Our Bubble app became slow with 1,000+ concurrent users. Database queries weren’t optimisable. Page loads hit 5+ seconds.

Custom UI/UX

No-code gives you templates. Customisation is limited.

Our marketplace looked generic. Competitors with custom apps looked polished. The difference mattered for trust.

Integrations beyond basics

Zapier handles common integrations. Custom APIs require workarounds.

We needed integrations with niche tools. Hours of workaround for what would be 10 lines of code.

Security requirements

Enterprise customers asked about security. “It’s on Bubble” wasn’t a satisfying answer.

No-code platforms are probably fine. But you can’t audit them. You can’t guarantee them.

The Cost Reality

No-code costs (using our Bubble project):

• Platform: $32/month
• Plugins: $50/month
• Workaround tools: $30/month
• My time configuring: 20 hours/month
• Total: ~$2,500/year + significant founder time

Custom code costs (replacing the same project):

• Development: $35,000 one-time
• Hosting: $100/month
• Maintenance: 5 hours/month
• Total: ~$36,200 year one, ~$2,500/year ongoing

Break-even: About 18 months.

If you’re not sure the project will exist in 18 months, no-code wins. If you’re confident it will, custom wins.

My Decision Framework

Use no-code when:

• Internal tool for your team
• MVP to test an idea
• Simple logic and data
• You’re not sure it’ll survive
• Speed matters more than scalability

Use custom code when:

• Customer-facing product that needs polish
• Complex business logic
• Scale beyond hundreds of users
• Security or compliance requirements
• You’re confident in the concept

The middle ground: Start no-code. Migrate when you hit limits.

This is what we did. Validated with no-code. Rebuilt with developers when it mattered.

The migration cost was real but manageable. The validation saved us from building the wrong thing.

The Honest Assessment

No-code is a tool. Like any tool, it’s right for some jobs and wrong for others.

The hype says you don’t need developers. The reality: you don’t need developers yet.

Eventually, if you succeed, you’ll outgrow no-code. Plan for that. Don’t get stuck.