What a website costs in Malaysia, and what the cheap one costs later

Ask what a website costs in Malaysia and the honest answer spans RM 1,500 to RM 100,000, which is useless as an answer and revealing as a fact. The spread exists because the word covers four different products that share a name: a page, a brochure, a marketing system, and an application.

A page is one screen with your phone number on it. A brochure is five to ten pages of who-we-are. A marketing system is a brochure plus the machinery that makes it findable and measurable — technical SEO, structured data, a second language, analytics, a content model somebody can actually update. An application is a site that does something: books, quotes, calculates, takes payment, talks to your back office.

Almost every disappointing website purchase in this country is somebody who paid brochure money and expected a marketing system, or paid marketing-system money and received a brochure. The ranges below are worth reading, but the sorting question comes first: which of the four are you buying?

Figures published by Xwork in August 2026. They are indicative bands for the Malaysian market, not quotes, and they describe what providers charge rather than what any particular site is worth.

The in-house row is on here because it is the comparison nobody runs. A junior developer at RM 3,500 a month is RM 42,000 a year before benefits, and Xwork suggests multiplying stated salaries by 1.3 to 1.5 for the true cost. That is a full custom agency build every year, forever, for one person who also needs managing.

Walk Production publishes the bands by page count instead, which is a more useful cut if you already know roughly how big the thing is. Their figures, published April 2026 and updated in August: a starter or campaign site of one to five pages at RM 1,500 to 5,000 budget, RM 3,000 to 8,000 mid-range, RM 6,000 to 12,000 premium. A corporate site of ten to fifteen pages at RM 5,000 to 9,000, RM 8,500 to 15,000, or RM 15,000 to 25,000 across the same three tiers.

Above that they list listed-company and MNC sites of twenty to thirty pages from RM 18,000 to RM 60,000, and e-commerce from RM 6,000 to RM 60,000. They describe these as indicative envelopes for planning rather than quotes, which is the right disclaimer and worth repeating.

Notice what happens when you put the two publishers side by side. Xwork's mid-tier agency band and Walk Production's premium corporate band cover the same money for arguably the same site. Page count and provider type are two different proxies for the same underlying thing, and neither is the thing itself.

Page count is the weakest driver on the list and the one every quote is built from, because it is countable. Five near-identical service pages generated from one template cost almost nothing more than one. Five pages that each need their own layout, photography and argument cost five times as much. The unit that matters is distinct designs, not URLs.

The first real driver is how much of it is bespoke. A theme with your logo dropped in is a day. A layout designed for your content is weeks. Both are legitimate purchases; only one of them will look like anything in three years.

The second is content, and it is the one that kills timelines rather than budgets. Somebody has to write the copy and somebody has to take the photographs. If that somebody is you, the project is cheaper and will take four months. If it is the studio, the quote goes up and the site ships. Decide this at quoting time, in writing, because a build waiting on content is the single most common reason a Malaysian SME website sits half-finished for a year.

The third is language. A properly bilingual site is not a translate button — it is a second set of URLs, a second set of metadata, correct hreflang, and a Malay version that reads like it was written rather than run through a machine. It is close to a second site's worth of content work and a fraction of a second site's worth of engineering.

The fourth is integration. WhatsApp, payments, a booking calendar, a CRM, an inventory feed. Each one is a separate system with its own failure modes, and each one is where a fixed-price quote turns into a variation order. The fifth is ownership: whether the code, the domain, the hosting and the analytics end up in accounts with your name on them. That one does not change the build cost at all and changes everything about what the site is worth.

A RM 1,500 site is not a scam and is sometimes exactly right. What it does is move costs from now to later, and it is worth naming the specific ones rather than gesturing at quality.

The site lives in somebody else's account. Builder platforms and a good number of freelancers set up the hosting, the domain and the analytics under their own login. When you part ways, you find out. Getting a domain moved out of an unresponsive developer's registrar account is a documented weeks-long exercise in Malaysian SME life, and the version where the developer has stopped replying is worse.

The plugins rot. A template build is typically a stack of third-party plugins whose combined maintenance is nobody's job. It works on the day it launches. Eighteen months later something has auto-updated, the contact form has stopped emailing, and nobody notices for six weeks because nobody was watching the enquiries.

It is slow on a phone on mobile data, which is where the customer is. Heavy themes carry the weight of every feature they support, including the ones you do not use. And it is usually invisible to a search engine in the ways that matter: no structured data, no clean sitemap, dynamic pages that return a 200 for a URL that does not exist, no Malay version, and a Google Business Profile that nobody set up because the brief said website.

The honest arithmetic on a cheap build is therefore not RM 1,500 against RM 8,500. It is RM 1,500 now plus a rebuild in year two, against RM 8,500 now and a site that is still yours in year five. Sometimes the first is still the right call — a business testing whether a service has a market should not spend RM 8,500 finding out. Just do it knowingly.

The line item that gets left out of comparisons. Published prices where a published price exists; the retainer figures are Xwork's August 2026 bands.

The infrastructure numbers are worth staring at, because they are the reason a modern static site is cheap to keep alive. A marketing site that gets ordinary traffic runs on free tiers for a long time, and the paid tiers are USD 20 and USD 25 a month rather than a hosting package sold at a markup.

We publish three tiers in ringgit and they map onto the ranges above rather than hiding from them. RM 2,500 is a first site: a small custom build for a business that needs to exist properly online, be fast on a phone, and get a WhatsApp message. It sits inside the freelancer band and is not a template.

RM 8,500 is a marketing build: up to fifteen pages, technical SEO, dual-language, CMS-backed sections, monthly reporting. That is the tier most businesses on this page actually need, and it sits below Xwork's mid-tier agency band while doing what that band describes. RM 25,000 and above is a system rather than a site — the tier where the website is a front door onto software.

These are fixed prices, not opening positions, and the reason we can hold them is scope discipline rather than corner-cutting. A quote that starts low and grows through variation orders is more expensive than a quote that was honest on day one, and it is the most common bad experience in this market.

Klang Valley agencies quote Klang Valley overheads, and for a Seremban business that difference is not a rounding error. It is also not the main reason to buy locally. The main reason is that the deliverable is intangible and the person who has to update it works in your office.

Being able to sit down together means the photographs are of your actual premises, the copy is checked by the person who answers the phone, and the handover happens in a room rather than in an email nobody reads. For a service business here, the sequence that produces an enquiry is a phone search, the map results, one glance at whether the business looks real, and a WhatsApp message. That is mostly won on your Google Business Profile, which is why we start there rather than on the homepage.

The Seremban and Negeri Sembilan pages go into what that looks like locally, including why the local search results for this trade are held by directories and job boards rather than by studios.