A No-Code Developer Built an App with Bubble: The Full Story

Short answer: A first Bubble app takes around 11 weeks of part-time work. Paid plans start near $29 a month and continue as long as the app is live. No-code holds up well for booking systems and internal tools, and starts struggling once data volume or calculation complexity grows.

---

Sahil's first Bubble app took eleven weeks. Five of those weeks were spent rebuilding the database, because the first time around he put everything into a single table.

He told me that with a laugh, but there's a real lesson buried in it. No-code doesn't mean you stop thinking. It means you stop memorising syntax.

Everything else stays. How the data is organised, what connects to what, what happens when a user clicks something. All of that is still yours to decide, and getting it wrong costs the same as it always did.

---

What the app was

An admissions and attendance system for a local tuition centre. Nothing exciting.

They had around 140 students tracked across paper registers and one Excel file that had been copied so many times nobody was sure which version was current. Working out monthly fees took their admin roughly two days, and mistakes happened often enough that parents had learned to check.

Sahil's brother taught there. The owner asked, half seriously, what software like that costs. A development agency had quoted Rs 1.8 lakh, which was never going to happen.

Sahil said he'd look into it. At that point everything he knew about Bubble was that it was a tool where you build apps without writing code.

---

Three weeks of pure confusion

Bubble's interface is genuinely overwhelming on day one. Elements on the left, a properties panel on the right, tabs across the top, and an entire separate section called workflows that seems to have its own logic.

Week one he worked through the official tutorials, which are fine. Week two he watched a YouTube course where someone built an Airbnb clone, and that was where the pieces connected for him. Not because Airbnb mattered, but because watching someone wire a list to a database and then filter it showed him the shape of the thing.

Week three he started his own app and got it comprehensively wrong.

He created one data type called Student and put everything in it. Name, phone, batch, fees, each month's payment, attendance records. All in one place.

That's spreadsheet thinking. In Excel a row is a student and you add columns forever. Databases don't work like that. Attendance is a repeating event, so it needs to be its own data type linked back to the student, otherwise you end up with fields called attendance_jan, attendance_feb, and a design that collapses in month thirteen.

When that finally landed, he rebuilt from scratch. Five weeks.

And honestly, those five weeks were the actual education. Everything after them went faster because of them.

---

The part he found hardest

Workflows.

Every action in Bubble is a workflow. Button clicked, form submitted, page loaded. You build them as visual blocks.

The concept is easy. The difficulty arrives with conditions. Show a reminder if the student's fees are pending, and their batch is the morning one, and today is past the fifth of the month.

He told me one of his workflows got long enough that he screenshotted it and printed it out to understand what he'd built. Which is funny, and also a fairly clear signal that something has gone wrong structurally.

He later split it into three smaller ones. That's the lesson most no-code builders learn late: keep each workflow doing one thing.

---

The money, both directions

The client paid Rs 34,000. He priced it deliberately low because he wanted a portfolio piece more than he wanted the margin.

Costs against that: the Bubble paid plan at roughly $29 a month, about Rs 2,450 at the time. A domain at Rs 760 a year. And a plugin for sending SMS reminders, billed separately.

Here's the part he handled badly. He never clearly explained the recurring cost. The client understood Rs 34,000 as the total price of the software, full stop.

Four months later the renewal came up and the conversation was uncomfortable. The client's line was that you buy software once, don't you?

He sorted it out by transferring the plan into the client's own account so they pay it directly, which is the right structure anyway. But he's clear that it was his mistake, not the client's misunderstanding. Every quote he writes now lists recurring costs separately, in bold, on the first page.

---

Where no-code gave out

The app ran fine while the data was small.

About a year in, once attendance records crossed roughly 4,000, the search page started taking four to five seconds to load. Users assumed it had frozen and clicked repeatedly, which made it worse.

He fixed most of it: pagination, and running searches with constraints applied at the database instead of loading the full list and filtering it in the browser. That recovered the speed and it has been stable since.

But he was honest about the ceiling. If the centre grew to 500 students with three years of history behind it, he doesn't think this approach survives without a real rethink.

That's the genuine limit. Small and medium scale, no-code works. Large data volumes and heavy calculation, it starts fighting you.

Actually, that's not quite the whole picture. Plenty of substantial products run on no-code platforms. But the people running them treat optimisation as a serious discipline, and at that level you have effectively become a developer working in a different tool, which rather defeats the original pitch.

If you want the coded route for comparison, the Flutter first app experience covers a different path, and what Play Store publishing actually returns is worth reading before anyone assumes app income is passive.

---

What came after

That one app produced three more clients, all through referrals. An appointment system for a dental clinic, an inventory tracker for a small manufacturing unit, and a volunteer management tool for an NGO.

The rates moved: Rs 34,000, then Rs 52,000, then Rs 71,000, then Rs 64,000 for a smaller scope.

He now also charges maintenance retainers, roughly Rs 4,000 to Rs 7,000 per client per month, covering small changes and support. That's the income he describes as the steadiest, because projects end and retainers don't.

One detail I keep thinking about: not one of his clients has ever asked what the apps are built with. Not one. What they care about is that the monthly fee calculation went from two days to about twenty minutes.

That's worth sitting with if you assume nobody will take you seriously without a computer science degree or real code behind you. Clients are buying a problem being solved. The method is your concern, not theirs.

He still wonders whether he should learn to code properly. He's started twice and stopped both times, because the hours he'd spend on JavaScript are hours another client project could fill. The logic holds up, and he also suspects that at some point it won't, and that he'd rather find out