They forwarded a rebuild quote the week the round closed. When should you leave no-code for a custom product? No-code means Bubble, Webflow, the editors you click together. You don't get a repo. Someone still sent an invoice from Bubble that afternoon. I told them to stay. They wanted the deck to say they were off it.
I take the rebuild as SaaS and web productswhen Bubble is what's stuck. I still send most of these back. If nobody logs in, stop, that's a website or a web product question, and a Webflow brochure is a website. How much a first version costs is the number after you already have to leave. I write a brief before I quote.
How to tell you should leave no-code for custom
I opened the app. A person had finished the job that morning. The rebuild was a PDF. Stay until you can point at what's broken in the live app this month.
Stay on no-code
The job still finishes in Bubble
- Someone finished the job in Bubble this week
- You changed a screen without filing a ticket
- What's annoying is not what's actually broken
Leave for custom
Bubble is what's broken
- Search still reads every row after you tried to fix it
- A person copies what the plugin can't
- A contract wants a repo, or a country the plan doesn't have
- The specialist and the plan cost more this year than a build
The cut
DelightLoophad to turn a campaign into a gift. Clicking that together meant plugins and a person sitting in the gap. I didn't. It's in selected work. A Bubble CRM with a form and a Stripe plugin is not that job. If people already pay you inside Bubble, don't start a second app next to it. That's a takeover of a live product.
What people mean by no-code or a custom product
The email says "we've outgrown no-code." On this one they meant the round closed. Other weeks it means a developer spent an hour and said Bubble doesn't scale. Or the search is slow, and nobody has opened the search.
If you're on WordPress, the fight is the theme. That's keep WordPress or rebuild. Shopify is checkout you didn't write, Shopify or a custom store. HubSpot already is the product, which is build custom software or buy a SaaS tool. Bubble is the one you built and still don't run. You click the screens. They host it, and when you ask for the app they hand you a spreadsheet.
I sell the rebuild. I have turned down that invoice. Custom is a repo you can deploy without their login. Then someone has to keep it up. Leaving because nobody owns the Bubble editor just gets you files you also don't own. Name who deploys before you change stacks.
What actually makes you leave no-code
I don't leave because Bubble looks cheap on a slide. I leave when the builder is what's broken. What belongs in a first version is the cut after you already have to go. How long that takes follows the job, not the editor you're leaving.
Search takes eight seconds
Eight seconds to open a record. Ask whether that's a fat image or a search that reads every row. I can tell from the live URL. If I can cut the search, that's days. If every screen is late because that's how Bubble fetches, you leave. Someone sent me a rebuild quoted as the only way to a three-second page. The three-second page was an index nobody added. A slow public site is already a product problem. The wait is still the problem. You're just in Bubble.
A person is the integration
There's a plugin that almost talks to the courier. A person copies the rest into a sheet. That person is now the product. I can connect Stripe when Stripe already charges the card. I can't make a plugin hold a rule it doesn't have. When the missing rule is why people pay you, you leave. When the plugin already takes the money, you stay, and you stop hiring me to rebuild the plugin.
An auditor asked for a repo
A customer sent a contract that wants the data in a country Bubble's plan doesn't offer. Or an auditor asked to read the source without a Bubble login. That's a leave. A blog post about SOC 2 is not. Don't rebuild because you think you will need it later. Rebuild when a contract is already on the table.
The Bubble bill passed the build
I opened an account where the plan, three plugins, and the Bubble person cost more in a year than a first version. Then you leave because of money. Not because the editor is ugly. I want those invoices before I quote. Tired of clicking is not a brief. The number still sits on how much a first version costs, and it only applies once Bubble is actually stuck.
When leaving no-code is the wrong next step
People leave because they raised. Or they want it to look like Linear. I don't rebuild for either. If nobody logs in, you wanted a business website. If HubSpot already does the job, you wanted HubSpot, and you never needed Bubble.
You raised
Investors like a repo. They also like invoices that still go out. Starting a rebuild the week the money lands is how you spend the round on a second copy of a job that already ran. Stay until something in the live app is broken. Then you have users, which means a takeover, not a blank first version. Name who owns it after launch before anyone starts the second app.
We'll need custom later
Later is not a brief. You'll pay to learn the job in Bubble. You'll pay again to rebuild. Fine, if Bubble is how you watch someone try the job. It is not a reason to rebuild on Tuesday while invoices still go out. I price the rebuild as a rebuild. The demo does not knock anything off.
An afternoon would do it
Search. A plugin that does nothing. A page that loads every record. If the rest of the app still does the job, that's a pass, not a quarter. I see rebuilds quoted so nobody has to do the boring pass. Do the pass. Then send the URL if something is still actually broken.
I watched a round get spent on a second Bubble. The first one still took the bookings.
Send the live Bubble URL and the custom quote. I'll point at what's broken before I quote a leave. Tell me what's actually stuck.
