WordPress is the right choice for most businesses, and the argument was never really about WordPress. What makes a site heavy, brittle and expensive to maintain is the layer bolted on top of it: the multipurpose theme, the page builder, the long plugin list.
Our own site runs on WordPress. Custom theme, no page builder, and on 15 August 2026 its homepage weighed 151 KiB on mobile. Same WordPress that produces four megabytes elsewhere.
Three different things sold under one name
Ask ten people for a quote and you will be sold three genuinely different products with the same label.
| Theme plus page builder | Theme with customisation | Custom theme | |
|---|---|---|---|
| Delivery time | Days | Two to four weeks | Four to ten weeks |
| Speed ceiling | Low, because of generated markup | Middling | Whatever you write |
| What you depend on | Theme and builder, both | The theme | The team that built it |
| Changing it later | You, within what the tool allows | Usually the same designer | Requires code |
| Year two cost | Licence renewals | Theme renewal | Maintenance |
None of them is inherently wrong. The failure is buying one and expecting another.
The number that takes this out of the realm of taste
On 15 August 2026 we ran Lighthouse 12.8.2 on several Persian pages, mobile form factor, identical settings, one cold run each. The page ranking first that day for the Persian query meaning "website design with WordPress" weighed 4402 KiB and scored 91. In the same round, a lighter page at 2972 KiB scored 48.
Server response made the difference: 90 ms against 1800 ms. A fast server can hide a four megabyte page in the lab. On a real phone, on a network that is not always steady, it cannot.
So there is a practical question for anyone about to commission a site: ask each bidder how many kilobytes the delivered page will weigh. The answer tells you more about what they build than any portfolio does.
Where the weight comes from
Images first. A three slide hero without compression outweighs our entire site on its own, and it is the one item the client usually adds personally, after handover.
Builder output second. A page builder loads its own CSS and JavaScript per block, even when the block holds a paragraph of text. The code is not bad; it is generic, and generic code has to cover every case.
Plugins third. Every active plugin loads on every page, including pages it has nothing to do with. A site with twenty five plugins pays that cost twenty five times.
A custom theme handles all three differently. We inject CSS into the HTML so nothing blocks rendering, subset the fonts, and removed animations entirely. None of that is inventive work. It is simply not available through a builder.
When WordPress is the wrong answer
We have watched it fail three ways. When the product is the software itself and users spend hours inside it after logging in, WordPress is a publishing system carrying an application it was not designed for.
When the business logic is genuinely unusual: dynamic pricing with ten variables, multi warehouse stock with reservation rules. You can write these inside WordPress, but you are building a framework that WordPress will not help you with.
And when you have an in-house team that is happier with something else. The worst sites are the ones built on a technology nobody in the organisation likes.
And when hiring an agency is a waste of money
Agencies do not usually write this down. We do, because it saves two months of argument later.
If your site is ten pages of information plus a contact form, has no custom process, and is never going to take a hundred thousand visits a month, a custom theme does not pay for itself. Buy a good theme, pay someone two days to configure it properly, and put the rest of the budget into product photography and copy. Your site will outperform plenty of far more expensive ones.
Come to a custom build when one of these is true: your traffic is large enough that speed moves revenue, you have a process no theme contains, or your brand is defined enough that looking like everyone else costs you something.
Questions that come up while collecting quotes
What does a WordPress site cost
There is no fixed figure, and anyone quoting one without seeing your requirements is leaving something out. We break the components down in an honest split of what building a site actually costs.
How long does it take
Days with an unmodified theme. Usually four to ten weeks for a custom build. The stages are laid out in the real order a site gets built in.
Can I migrate from a theme to a custom build later
Yes, and starting custom is usually cheaper than migrating, because content laid out inside a page builder does not travel to another theme. If you already know you will go custom next year, invest less in the builder this year.
Where to go next
If you have settled on WordPress, infrastructure is the next decision, and we list the criteria in what to look for in hosting built for WordPress. If you want to know what your future site's speed will actually hang on before you commission it, the share hosting takes in your Core Web Vitals covers it.
When you reach the quoting stage, a WordPress build on a custom theme is where we work. If your brief shows that an off the shelf theme is enough, we will say so and decline the project.
Sources
- How the Lighthouse performance score is calculated for the weight of each metric
- The Web Vitals reference page for the LCP, INP and CLS thresholds
Measurements were taken on 15 August 2026 and the tested pages came from live Google results the same day.
Comments & Questions
Have a question about this article? Ask, we'll answer.
No comments yet; be the first.