No-Code, Off-the-Shelf, or Custom Software: The Cost of Growth

The question usually arrives with a number attached. "Squarespace is thirty dollars a month. Why would I pay you thousands to build the same thing?"

Fair question. It usually gets a poor answer from people in my trade. "Custom is more scalable" is a slogan, and "you'll own it" is true without explaining anything on its own.

So let me start where most articles like this one refuse to. For most businesses, most of the time, the ready-made tool is the right answer, and I'll say it again later with the cases attached.

What the ready-made tools don't advertise is what the bill looks like in year three. They're cheap at the start. They get expensive later, in a way you can see coming if you know where the meter is.

Every figure below links to the page it came from, and almost all of those pages belong to the platforms themselves: their pricing tables, their fee schedules, their export documentation. Nothing here rests on a comparison blog or a number somebody repeated without a source. Where a widely quoted statistic didn't survive a check, I've left it out and said why.

The three things I'm comparing

People use these three words interchangeably in the same conversation, usually while comparing quotes, and that's where a fair number of bad decisions start. This is what each one means below.

No-code builders

Most people call these website builders, which is the honest name for the job they do. You assemble the site or the app yourself, in a visual editor, by dragging things onto a page and filling in settings. There's no code involved on your side, and there's no server for you to think about. The platform hosts it, runs it, updates it and keeps it secure, and the whole thing lives inside their account.

Examples: Wix, Squarespace, Webflow and Bubble. The first three are mostly for websites. Bubble goes further and lets you build an application with logic and a database, still without writing code.

You're trading the machinery for speed, and for a lot of businesses that's a good trade.

Off-the-shelf platforms

Software somebody already built for a job that thousands of businesses share, which you then configure to fit yours. You're not designing how it works. You're choosing settings, picking a theme, and adding paid extensions when the standard version doesn't reach far enough.

Examples: Shopify for online stores, WordPress for content sites, Airtable for structured data and internal coordination, and Zapier for connecting the rest of them together.

This is the widest category and the messiest one. WordPress can be run on your own hosting with your own code in it, which puts it halfway into the next category. Shopify sits at the other end, where the store is yours and the checkout is firmly theirs.

Custom software

Software written for your business, by a developer, against your requirements. The code belongs to you, the database belongs to you, and it runs on hosting in your name. No part of it waits on a vendor's roadmap.

It costs more at the start, and it comes with jobs somebody has to do: hosting, backups, security updates, and answering the phone when it breaks. One example is a digital menu and ordering platform I built for a client. Custom was the right call there because what happens in a café during service was the product, and no off-the-shelf tool described it.

The lines between the three blur, and I'll say so where it matters. The part that matters more is that they don't get expensive for the same reasons.

Start with the part that argues against me

WordPress runs 40.7% of all websites, and 58.9% of sites whose content management system is known, according to W3Techs on the day I'm writing this. That page updates daily, so the number will have moved slightly by the time you read it. The direction won't have.

That's not a story about people settling for something. Ready-made tools won because they're good, and because what they replaced was worse for nearly everyone. A business owner can have a real site up this weekend, with hosting, TLS, backups, a content editor and a payment button, for the price of two coffees a week. Twenty years ago the same thing took a developer, a server and a month.

So the case for custom software can't be "builders are bad," and that's not the case I'm making. Mine is narrower. The platform price and the custom price grow for different reasons, and the point where those two lines cross arrives later than platform sellers imply and earlier than most owners notice.

The meter, and who holds it

The whole argument fits in one sentence, and the rest of this article is evidence for it.

A platform bill is indexed to your success. A custom system's bill is indexed to your requirements.

Ready-made tools charge you more as you get bigger. That's how a $30 customer becomes a $3,000 customer without the vendor selling anything new. Custom software charges you when you ask it for something it doesn't already do.

Three meters do almost all the work here, and it pays to know which ones you're wired into. Most businesses are wired into more than one and only ever notice the first.

Meter one: per seat

Airtable's published pricing is $20 per user per month on the Team plan and $45 per user per month on Business, both billed annually.

For three people that's $60 a month, which is a bargain. Nobody should build software to avoid paying $60 a month.

Now put twenty-five people on it, which is a normal number for a growing business once the warehouse and the office both need to see the same records. Business is then $1,125 a month, or $13,500 a year. The part that stings is that most of those twenty-five aren't power users. They open one view, change one field, and close it again. You're paying a full seat for four minutes a day.

Per-seat pricing is fair to the vendor and blunt for you. It bills headcount, and headcount can't tell the difference between somebody who lives in the tool and somebody who visits it.

Meter two: per usage

Zapier's pricing page lays the usage meter out more plainly than most. On the Professional plan, billed annually: 750 tasks a month is $19.99, 10,000 tasks is $129, 100,000 tasks is $489, and 2,000,000 tasks is $3,389 a month.

That ladder is a fair deal and a warning at the same time. The fair part is that the price per task drops hard as you climb, so nobody's being gouged at the top.

The warning is in what counts as a task. It's one action in one automation, so a single order that fires a confirmation email, updates a spreadsheet, creates an invoice and posts to a staff channel spends four of them. Triple your orders and you've tripled the meter. Nobody decided to spend more money. The business just got busier.

At $19.99 a month you'll never look at it. At $489 a month you're spending $5,868 a year on glue between systems you already pay for separately.

Meter three: per transaction

This is the expensive meter, and people agree to it faster than any other, because a percentage doesn't feel like a price.

Squarespace's help page on transaction fees lists them by plan. On the Basic plan, store orders carry a 2% fee and digital products and memberships carry 7%. Core is 5% on digital products, Plus is 1%, and Advanced is 0%. Older plans are still out there with their own rates, and plans without a digital products entitlement sit at 9%.

Those fees sit on top of what Stripe or PayPal charge you to move the money. They're the platform's cut for hosting the transaction.

Shopify does the same thing in a different shape. Its pricing page charges a third-party transaction fee when you use a payment provider other than Shopify Payments: 2% on Basic, 1% on Grow, 0.6% on Advanced, and 0.2% on Plus. Shopify's own help center is explicit that "transaction fees are in addition to processing fees charged by your third-party payment provider".

Percentages hide their size, so put one in money. Sell $3,000 a month in digital products on Squarespace Basic and 7% is $210 a month, which is $2,520 a year on top of the subscription and on top of what Stripe already took. Sell $10,000 a month and it's $8,400 a year. Move up to the plan that charges 0% and you pay a bigger subscription instead, which is the whole reason the upgrade path is built that way. The fee is what makes the expensive plan look cheap.

None of this is a scam. It's a rental agreement where the rent is a slice of your revenue, and while the revenue is small it's the best deal in software. Once the revenue is large you're handing over a percentage of your business every month, and you keep handing it over for as long as you stay.

The price is theirs to change

There's a second cost that never appears in a comparison table, and it catches the people who did their homework properly at the start.

You priced the decision once. The platform gets to price it again whenever it likes.

In April 2023, Bubble announced a new pricing model built on "workload units," which replaced the old server capacity model. Bubble's own description is that "workload measures the work Bubble does to power your app." New plans went live on 1 May 2023. Apps on legacy plans could stay on them with a price increase of about 10%, and that grandfathering ended on 1 October 2024.

Sit with that for a second if you had a business running on it. The unit your bill was counted in changed. Not the rate, the unit. Every sum you'd done about what the app would cost at three times the size was now written in a measurement that had stopped existing, and you had eighteen months to redo them all.

Webflow ran a gentler version of the same play this year. In May 2026 it merged the CMS and Business site plans into a single Premium plan, with limits that match neither of the old ones. Anyone who picked Webflow by weighing CMS against Business was comparing two things that are now gone.

I'm not calling either company dishonest. This is simply what SaaS is. Vendors reprice, merge tiers, retire plans and change the meter, and they're entitled to do all of it. The part that concerns you is smaller and harder: on a platform, one line of your cost of goods sold is a number somebody else revises, and the only vote you hold is to leave. Whether leaving is realistic is the next section, which is the one nobody reads until it's urgent.

The statistics I'm not going to use

Two numbers dominate this topic. Both would help me. I'm using neither.

"By 2025, 70% of new applications will use low-code or no-code technologies." You'll see this quoted in almost every article about no-code, usually with Gartner's name attached and no link. Gartner does publish free press releases on this market, including one titled "Gartner Forecasts Worldwide Low-Code Development Technologies Market to Grow 20% in 2023", but the research behind the 70% figure sits behind a paywall that almost nobody quoting it has been through. And look at what it is: a forecast about enterprises, published years before the date it describes, aimed at companies several sizes larger than the ones asking me this question. 2025 came and went without anybody going back to mark it. A prediction nobody scored afterward isn't evidence of anything.

Any "X times faster" or "X% cheaper" figure from a platform comparison guide. Try this one yourself. Search for the pricing of any major builder and look at who wrote the first page of results. When I searched for Webflow's 2026 pricing, most of what came back was written by agencies that build Webflow sites. Their numbers are usually right. They're also not neutral, and neither am I, which is why every link above goes to the platform's own page instead of to somebody's summary, mine included.

There's a selfish reason for being this careful as well as an honest one. If I quoted a flattering number and it turned out to be junk, you'd be right to wonder how carefully I check anything else.

The cost nobody prices: leaving

Every comparison of website builders against custom software argues about the monthly fee, and the monthly fee is the small number. The big one is what it costs to stop. You can look it up this afternoon, for free, on the help pages of whatever you're using now.

Read a few of them together and the same pattern shows up every time: your records leave, your system doesn't.

Squarespace publishes exactly what its export includes. Layout pages, one blog page with its posts, text blocks and image blocks come out as a WordPress XML file. What doesn't come out: other page types including store pages, album pages, index pages and portfolio pages, additional blog pages, audio blocks, video blocks, product blocks, drafts, style settings and custom CSS. Your design isn't in that file anywhere. The same page adds that "it's not possible to export content from one Squarespace site and import it into another," which is worth reading twice.

Wix is the most direct of the group. Its help center explains that a Wix site is built with Wix's technology and has to be hosted and operated on Wix's servers, so there's no export of the site to another host at all. You can export a CMS collection as a CSV file, so the records are yours. The site isn't.

Webflow does let you export code, and it publishes what the export leaves behind: CMS content, ecommerce, user accounts, form handling and localization all stay put. What you get is the HTML and CSS of a site with its moving parts switched off, plus CSV backups of your collections. Useful, and a long way from a working website.

Shopify comes out of this best, and deserves the credit. You can export products, customers and orders as CSV, and the documentation is straight about the edges, including that "the images associated with each product aren't included in the CSV file." Then there's the rest of it. Your theme is written in Shopify's template language, your apps are Shopify apps, and your checkout is Shopify's checkout. The catalog moves. The store stays.

Custom software hands you a code repository, a database you can dump and a domain, all of which will run somewhere else the day you decide they should. That's the real argument for building, and it's an argument about insurance rather than quality.

The reason this outweighs the subscription is timing. You pay the exit cost at the worst possible moment, which is the month the platform reprices, or the month your business finally needs the one thing the platform won't do. That's when you find out the rebuild is a project rather than an afternoon, and that you'll be paying for two systems while it runs.

Nobody budgets for it at the start, because at the start nobody's leaving.

When to move off a no-code platform

Most people expect the ceiling to be about size, as though the platform gives out once you get big. That's rarely how it happens. Plenty of large businesses run happily on Shopify and always will.

The ceiling is about shape. It arrives when what your business does stops matching what the platform can describe, and it usually announces itself in one of these ways:

• Your checkout has to be different. Deposits, part payments, quotes that become orders, per-customer pricing, an approval before payment. Hosted checkouts are deliberately rigid, because that rigidity is what keeps them secure and reliable.
• The workflow crosses systems that don't know about each other. An order has to reach the warehouse, the accountant's software and the courier, and it changes shape at every stop. This is where the automation bill from meter two starts climbing.
• You're buying seats for people who use one screen. Ten warehouse staff marking items picked, at full seat price, is a signal worth listening to.
• A limit that only enterprise sales can raise. Records, API calls, storage, bandwidth. Once a limit is negotiable instead of published, you've left the pricing page and walked into a sales process.
• You have to answer for the data. Retention rules, an audit trail, a regulator or an insurer with an opinion about where records live.
• The workarounds have quietly become the system. Three linked spreadsheets, somebody rekeying orders every morning, an automation that one person understands and nobody has documented. Everyone still calls it "the platform."

The last one deserves a number before you decide anything else. Two hours a day of somebody's time, at $25 an hour, runs to roughly $13,000 a year. That's almost always larger than the subscription you were worried about, and it stays invisible because it gets paid in staff hours instead of on a card.

When ready-made is the better answer

A comparison that only lists the other option's problems is an advertisement, so here's the case against custom software. For most people reading this page it's the stronger case.

You haven't proven the thing sells yet. This is the big one. Custom software is a bet on a business model you've already tested, so if you don't yet know what you're selling, to whom, at what price, the platform is exactly the right tool for where you are. Build the cheapest thing that can take money, and learn from it.

Your needs really are standard. A brochure site with a blog. A store with normal products and a normal checkout. A booking page with normal slots. Standard is a compliment here. Somebody has already built that better than a custom project would for $30 a month, and a company with a security team keeps it patched.

Your volume makes the percentage small. At $2,000 a month in sales, a 2% fee is $40. Nobody should spend thousands to save $480 a year. Percentages only turn into arguments at volume.

It's an internal tool for a handful of people. Three or four people coordinating work in Airtable or Notion is the cheapest software your business will ever run. Build nothing.

You have a hard deadline in weeks. You can be live on a platform this month. That fact alone decides plenty of these calls, and it should.

Nobody is going to look after custom software. This is the failure mode I get called into most often. A custom system needs hosting, backups, dependency and security updates, and a person who picks up when it breaks. With nobody holding that, a platform's dull little subscription is safer than software you own and can't maintain. I've written separately about what a website costs after it launches and about who should build it, including the times the answer is an agency.

One warning about my own side of the fence. Custom software built badly is worse than a platform in every way, and it stays worse. Owning something bad is still owning something bad.

A rule of thumb you can apply this week

You don't need a spreadsheet for this. You need four numbers, added up across three years, because a single year flatters the platform and five years flatters me.

1. The subscriptions. All of them, including the automation tool, the form tool, the scheduling tool and the plugins.
2. The percentage. Your annual sales through the platform, times the transaction fee. Use the fee on your current plan, from your platform's own fee page.
3. The workaround labor. Hours a week spent copying, rekeying and fixing, times a real hourly cost. Be honest here, because this is where the money actually is.
4. The exit cost, discounted for the fact that you might never pay it. What would rebuilding elsewhere cost if you had to?

Put that total next to a build cost plus three years of hosting and maintenance. If the platform wins, stay where you are and stop reading articles like this one. If it's close, also stay, because a close call means the ceiling isn't here yet. Move only when the platform loses clearly, and start moving before it turns urgent. A migration you planned costs about half a migration you were forced into.

No-code vs off-the-shelf vs custom software, side by side

Everything above in one place. Some rows favor me and some don't, which is the point.

	No-code builder (Wix, Squarespace, Bubble)	Off-the-shelf platform (Shopify, WordPress, Airtable)	Custom software
Time to live	Days	Days to weeks	Weeks to months
Cost at the start	Tens of dollars a month	Tens to hundreds a month	Thousands, once
What the bill grows with	Seats, usage, sales	Seats, sales, apps and add-ons	What you ask it to do
Transaction fees	Up to 7% on digital goods, by plan	0.2% to 2% on third-party payments, by plan	Card processing only
Who sets the price next year	The platform	The platform, plus each app vendor	You and your hosting bill
Fits an unusual workflow	Poorly	Partly, with apps and glue	Yes, that's the point
Who fixes a security hole	The platform, usually before you notice	The platform, plus you for plugins	You, or whoever you hired
What leaves with you	Records as CSV, no design or logic	Records, some templates, no checkout or apps	All of it: code, data, domain
Ongoing effort from you	Almost none	Some: updates, plugins, apps	Real: hosting, updates, a person to call
Best when	You're proving the idea, or your needs are standard	Your business is a normal shape at real volume	The shape of the work is the business

The bottom row is the honest summary. Read the three columns as three bets about how unusual your business is, rather than as three grades of the same product. Most businesses are less unusual than their owners believe, which is a good thing and saves everybody money.

What to do this month

Whichever way you're leaning, all of these are cheap and all of them pay off either way.

1. Find your platform's fee page and read it. Not a comparison blog, the platform's own page. You want to know the percentage you're paying and what it's charged on.
2. Add up every subscription in the stack. The builder, the automation tool, the plugins, the seats. Most owners are surprised by the total, and the surprise is the useful part.
3. Price the workarounds. Count the hours a week your team spends moving data by hand and multiply by a real rate. That number decides more of this question than the subscription does.
4. Read your export page before you need it. Every platform above publishes one. Find out what leaves with you while it's still an idle curiosity.
5. Run a test export today. Download the file and open it. A backup nobody has ever opened is a guess.
6. Keep the domain in your own name. Not the agency's, not sitting inside the platform's account. Yours. It's the highest-value hour on this list, because it's what makes every later move possible.

When you do hire somebody, ask what you own on the last day: domain, code or account, database, analytics, all of it, in your name. My answers are on the questions page, and how I run a project is written down as well.

Summary

No-code and off-the-shelf tools are cheap to start and expensive in a specific, predictable way. Their bill is attached to your success: seats as you hire, usage as you get busy, a percentage of every sale. Airtable runs $20 to $45 per user per month, Zapier moves from $19.99 to $3,389 a month on one plan depending on volume, and Squarespace takes up to 7% of digital product sales depending on your tier, all of it on top of card processing. Custom software bills on a different principle. It's attached to your requirements, and it goes up when you ask for something new.

The price is also theirs to change. Bubble replaced the unit its bills were counted in during 2023 and closed the old plans eighteen months later. Webflow merged two site plans into one this May. That's ordinary SaaS behavior, and it means one line of your cost base belongs to somebody else's pricing meeting.

Then there's the cost nobody prices, which is leaving. Read the export pages side by side and they agree with each other. Your records come out as CSV files. Your design, your checkout, your automations and your logic stay where they are. Squarespace states outright that its export can't even be imported into another Squarespace site. That bill lands at the worst possible time, which is the month you've outgrown the tool.

None of which adds up to "build custom software." If you haven't proven the idea, if your needs are standard, if your volume keeps the percentage small, or if nobody will maintain a system of your own, the platform is the right answer and I'll tell you so on a call. Ready-made tools are an excellent deal at the start of a business, and plenty of businesses never need to leave them.

Know which meter you're wired into. Price the workarounds honestly. Read your export page while it's still an idle question. Nobody gets hurt here by choosing a platform. People get hurt by never looking at the terms of the deal again after the day they signed up.