Builder or custom theme gets argued like a religious war and settled like a spreadsheet. Two boring questions decide it: who edits this site after launch, and does one layout repeat across a pile of items. Answer those and your budget band mostly picks itself.
Bias up front: we built client sites on Divi for years and wrote posts defending it, including a 2020 piece arguing its limitations weren’t really limitations. Both still up. We now ship custom themes on the block editor with ACF Pro, and we still build on builders when that’s the right call. Budget decision wearing a technical costume.
What are we actually comparing?
A page builder (Elementor, Divi, Bricks, Beaver Builder) hands your editor a visual canvas with dozens of controls on every element. A custom theme hands them a short list of purpose-built blocks that do what the site needs and nothing else.
For most small business work, “custom theme” means the native block editor plus a few fielded blocks built with ACF Pro. Fielded meaning labeled inputs. Headline. Subhead. Image. Button text. Fill in the boxes, the layout renders itself.
If the comparison you are weighing is Divi against a tool outside WordPress, I built the same site in Craft CMS and in Framer and wrote up which one I would hand to a marketing team.
Where do the AI site builders fit into this?
It comes up on every sales call now. There’s a category of tool where you describe your business in a sentence and something that looks like a website appears a minute later. If that’s new to you, you’re not behind. We use tools in that family daily, inside the build.
Honest read: what comes out is a starting point, not a finished business asset. Framed house. Walls up, nothing wired.
If you already tried one and couldn’t name why the result felt off, it’s almost always the same short list. Spacing that’s close but never consistent, so nothing lines up. Type that looks fine on your laptop and shoves the headline into four lines on a phone. Copy that describes a business instead of yours. Images that move the layout as the page loads. No form logic, no notification routing, no schema, nothing done for a screen reader. Those aren’t prompt problems and you weren’t one tweak away. That gap is the job, and it’s most of what you’re paying for when you hire us. Here’s what that work covers.
Question 1: Who edits this site, and how often?
If one non-technical person touches a few pages a few times a year, a builder’s freedom beats its overhead. If several people edit weekly, or the editor is nervous about breaking something, fielded blocks win. Far less to break.
Drag-and-drop freedom cuts both ways. Anyone can drag. Padding gets nudged. A column stacks like a ransom note on a phone and nobody notices for weeks, because the person who changed it only looks at their laptop. I used to think handing clients every control was generous. It isn’t. It’s handing back the design decisions they paid you to make, and handing you a Tuesday night you never budgeted for.
Question 2: Does one layout repeat across a lot of items?
If yes, custom blocks and custom post types win almost regardless of budget. Repetition is the cost driver clients underestimate most, and a builder makes you pay for it once per page, every design change.
Say you run a landscaping company with 14 service pages. In a builder you build the first, duplicate it 13 times, swap the copy. Fast, and fine. Then in March every service needs a “starting at” line under the headline. Open 14 pages. Next spring: 22 services and a new call-to-action design. Open 22 pages.
With a custom post type and a block, the layout lives in one template. Add the field once, all 22 pages get it. That’s why booking sites, stores, staff directories and multi-location businesses flip to custom much earlier than brochure sites.
What can native WordPress do now, with no builder at all?
More than most people assume. Global colors and type defined once and, if the theme switches off the custom color picker and font sizes, genuinely enforced. Synced patterns, so one edit updates every page using them. Template parts for headers and footers, the Query Loop block for listing any post type, and block locking so editors change words and images but not structure. In plain English, theme.json tells the editor which colors and sizes exist, so “make it bigger and purple” isn’t in the menu.
The four things still genuinely faster in a commercial builder
- Per-device fiddling without writing CSS. Different padding on tablet, smaller headline on mobile, hide this on desktop.
- Motion and interaction. Core gives you sticky positioning and a fixed background on Cover. Scroll animation, staged reveals and parallax are CSS or a plugin.
- Bundled widgets. Sliders, tabs, popups, mega menus. Core has a basic accordion, the rest are a plugin or a developer.
- One-off landing pages on a two-day turnaround. A builder gets there faster.
If your site lives on that list, a builder is the correct tool. That’s not us waving you off. It’s us quoting the builder build instead of the custom one. We sell both.
Page builder vs custom theme, by budget band
| Budget | Typical project | What we’d recommend |
|---|---|---|
| Under ~$2,500 | Brochure site, 5 to 10 pages, one occasional editor | A builder, or a fixed-price launch package |
| $2,500 to $5,500 | More pages and design, nothing that repeats | Either. Question 1 breaks the tie |
| $5,500 to $8,500 | Booking, online store, repeating records | Custom blocks on a custom theme |
| $9,500 to $12,000 | Membership, courses, gated content | Custom, builder off the table |
| $15,000+ | Custom platform with real application logic | Custom. A builder here is a liability |
Under about $2,500: a builder is a rational call
Our Launch package is $2,400 flat: brand kit, custom design, the full build, ten pages populated with real content, on-page SEO and training. At that number nobody is hand-authoring a dozen bespoke block types, and nobody needs to. Ten pages of brochure content doesn’t repeat enough for custom blocks to earn their keep. More on the small business websites page.
The band just above, $2,500 to $5,500, is the ambiguous one: more pages and design, still nothing repeating, so it comes back to who logs in. One careful owner? Builder. A rotating cast of staff? Fielded blocks. Same conversation either way, and the calculator on our pricing page will put you in a band in about two minutes.
$5,500 to $8,500: repeated layouts flip the math
Booking sites start at $5,500 with us, custom design and features at $6,500, an online store at $8,500. What those share is records: appointments, staff, products, locations, projects. Once a layout repeats, arithmetic takes over. How many times do you want to pay to change the same design? Custom blocks cost more on day one and less on every change after. Examples in our work.
$9,500 and up: the builder becomes the constraint
Membership builds start at $9,500 with us, courses and LMS at $12,000, custom platforms at $15,000. Up here a builder costs you three ways. Assets: every builder ships a base CSS and JavaScript payload on pages that don’t need it, and that lands on INP (Interaction to Next Paint, Google’s measure of how fast a page reacts when you tap it), not only page weight. Add-ons: each pack pins your upgrade path to a vendor you don’t control. Storage: content lives in the builder’s format rather than plain post content, which is why “we’ll just switch themes later” almost never is.
What do you give up going custom? Stated plainly.
- You can’t invent a new section shape yourself. A layout nobody built a block for is a developer ticket. Usually a small one. Still a ticket.
- Higher upfront cost. Obviously. Lower cost on every change after. That’s the trade.
- You need a developer who can read the code. The real lock-in worry, and manageable if the build uses boring standard parts: WordPress core, ACF Pro, Gravity Forms. An exotic stack nobody else uses is worse lock-in than any builder.
- Experiments are slower. Four landing page variants this month? Friction a builder wouldn’t add.
Either way, someone has to keep the thing patched, and this is the one line I’ll never help a client cut. Not backups, not updates, not security. An unpatched site doesn’t fail on your schedule. It fails the week you’re busiest, and the cleanup costs more than a year of care. Our Care plan is $300/month, month to month, no contract, and a site we didn’t build gets an audit and onboarding pass first. A builder site needs all of that plus keeping the builder and its add-ons current. Either route, there is no lock-in: the domain, design files and content stay in your name, and the site runs on standard WordPress.
Already on a builder? Five rules so it isn’t a rebuild later
- Global colors and fonts get set once, in the builder’s global styles. Never a hex code typed into a module. A brand blue living as six near-identical hex values has to be found and replaced by hand.
- Repeated sections go in the builder’s saved or global template feature, not copy-paste.
- Anything that is a list of things belongs in a custom post type. Services, staff, locations, projects. Fourteen hand-built pages is a decision you regret in year two.
- Custom CSS lives in one place. A child theme stylesheet, or the builder’s global CSS box. Not sprinkled across per-module fields nobody will find.
- No add-on module packs. Each one is another thing rebuilt by hand when you move. Same for forms: a standalone plugin (we use Gravity Forms) so yours survive a theme change.
Those are the rules we follow on builder builds, and the first things we audit on somebody else’s. We quote migrations between $700 and $1,800, and where yours lands tracks how much has to be rebuilt by hand rather than moved.
So why did we change our own position?
Not because the builder got worse. Because the work filled up with repetition, and freedom to make every page different is worth nothing when a stack of pages is supposed to be identical. We sell a $2,400 flat-rate launch and we sell custom block builds. Both are us. Pointing someone at the cheaper one is a different line on the same menu, not a nudge out the door.
Frequently asked questions
Is Elementor good enough for a business website?
Set up properly, yes, for a brochure site with one editor and no repeating content. It stops being right when the same layout repeats across many records, when several people edit weekly, or when the plugin’s overhead becomes the thing you’re fighting. The license is an annual cost you carry forever.
Do I need a custom WordPress theme?
You need one when the same layout repeats across a lot of items, or when your editors need constraining so they can’t break the design. If neither applies yet, a professionally built builder site does the job for less on day one. That’s our $2,400 Launch package, not a DIY weekend. Put the difference into copy and photography, which is where a small site usually wins or loses.
Is Gutenberg good enough on its own in 2026?
For plenty of small business sites, yes, once somebody has configured it. Global styles, patterns, template parts and query loops cover the structural work that used to need a builder. That configuration is the developer part: theme.json, locked blocks, fielded blocks. Per-device styling, scroll animation and sliders still mean a builder, a plugin, or us.
What happens to my content if I stop using my page builder?
Your words are still in the database. Your layout usually isn’t usable without the plugin. Divi 4 and earlier stored layouts as shortcodes, so deactivating it left bracketed tags all over the page. Divi 5 uses a structured data format, tidier to look at and just as unusable without the theme. Elementor keeps layout data in post meta rather than the post body. Native block content stores as HTML inside the post, the most portable of the three. Plan on rebuilding layout, not losing text.
How much does a custom WordPress theme cost?
With us, custom design and features start at $6,500, and the number moves with how much repeats and how much logic sits behind it. Full breakdown in our guide to what a small business website costs, plus a calculator on the pricing page.
The short version
This gets decided by whoever logs in on a Tuesday to change the hours, and by how many pages share a layout. Pick for the site’s second year. That’s when the bill for the wrong choice shows up.
Two minutes on the pricing calculator lands you in a band. Then send it over with two sentences: who logs in, and what repeats. Bring the current site, or the AI draft you gave up on. We’ll tell you which build we’d quote and why, before you’ve spent a dollar. Sometimes that’s the $2,400 builder package instead of a $6,500 custom one. We quote that one just as happily.