Skip to content

Web · · 9 min read

Web Application Development in Malaysia: What Businesses Should Plan First

A web application is not a website with a login. It is a business system — people enter data, review work, approve things, pull reports and coordinate — delivered through a browser. The Malaysian market for them is large and still growing, and the failure rate of software projects in general has been stubbornly high for thirty years. This guide puts the two sets of numbers side by side and draws out what a company should decide before anyone writes code.

The Malaysian context in numbers

The demand side is close to saturated. The Department of Statistics Malaysia’s 2024 ICT survey put household internet access at 96.8% — 98.8% in urban areas and 90.3% in rural ones — with 99.5% of households having a mobile phone and 97.9% of those being smartphones. The same programme’s open dataset shows that 99.4% of Malaysian internet users went online by mobile phone in 2025 and only 39.5% by portable computer. Any web application that staff or customers will use outside an office is, in practice, a mobile application that happens to run in a browser.

RM1.29 trillione-commerce income of Malaysian establishments in 2024, up 8.8%[3]
74.4%of Malaysian establishments had a web presence in 2024; 58.7% had their own website[3]
99.4%of Malaysian internet users go online by mobile phone (2025)[2]
47%of projects that miss their goals do so primarily through poor requirements (PMI)[6]

On the business side, DOSM’s Usage of ICT and E-Commerce by Establishment 2025 report (reference year 2024) found 97.2% of establishments using computers and 95.3% with internet access; 74.4% had an online presence, but that presence was mostly borrowed — 80.3% used social media, 58.7% had their own website, 44.1% sold through e-marketplaces. E-commerce income reached RM1,288.1 billion, up 8.8% from RM1,184.1 billion, and the business-to-business share was RM879.6 billion of it. The ICT and e-commerce sectors together contributed 23.5% of the national economy in 2023, according to DOSM’s Malaysia Digital Economy 2024 release as reported by MIDA. The gap between near-universal connectivity and a business base that still runs on social media pages and marketplaces is, bluntly, the market for custom web applications.

What the data says about why projects fail

There is no Malaysian software-failure dataset, so these are global figures, and they are old enough to be uncomfortable but consistent enough to be trusted. The Standish Group’s CHAOS Summary 2009 reported 32% of projects succeeding (on time, on budget, with the required features), 44% challenged and 24% failed, with the success rate between 2000 and 2008 never above 35%. Cost overruns on challenged projects ran at 54% and time overruns at 79%. Size was the strongest predictor: projects under US$750,000 in labour had a 71% chance of success, projects between US$750,000 and US$3 million a 38% chance, and projects over US$10 million a 2% chance of coming in on time and on budget. The same report quotes the finding that about 50% of specified software features are not used.

Standish CHAOS 2009: project outcome by labour cost (normalised for geography)
Labour cost (US$)Chance of success
Under 750,00071%
750,000 — 3 million38%
Over 10 million2%
[5]

The Project Management Institute’s 2014 Pulse of the Profession on requirements management adds the cause. When projects did not meet their original goals, inaccurate requirements management was the primary reason 47% of the time; "inaccurate requirements gathering" was named a primary cause of failure by 37% of organisations in 2014, up from 32% the year before; 5.1% of every dollar spent on projects was wasted through poor requirements — US$51 million per US$1 billion; and only 49% of organisations said they had the resources to manage requirements properly. BCG’s 2020 study of digital transformations, drawing on 70 companies and a survey of 825 executives, found 70% fell short of their objectives, and that the odds rose from 30% to 80% when six factors were all present — an integrated strategy with clear goals, leadership commitment, high-calibre talent, agile governance, monitoring of outcomes, and a business-led modular technology platform. Companies that addressed only three or four of the six failed.

Read together, the three studies say the same thing in different decades: small, well-specified, business-led projects succeed; large, vaguely specified, IT-led ones do not. Nothing in the Malaysian market changes that arithmetic.

Why requirements are the mechanism

A web application encodes a business process. If the process is not written down — who may create, see, edit, approve, export and delete each kind of record, and what happens when something is late, rejected or exceptional — the developers will write it down for you, in code, and the first time anyone discovers the gap is in user acceptance testing. That is the 47%.

The second mechanism is the one Standish measured: features. Half of what is specified is not used, which means half of the budget on a fully specified system buys nothing, and worse, each unused feature is a screen to maintain, a permission to reason about and a place for bugs to live. A first release that does one workflow end to end, in production, with real users, teaches more about the second release than any workshop.

The third is the mobile reality above. A system designed on a 27-inch monitor for a desk-bound administrator and then used by a technician on a 6-inch phone in a warehouse with one bar of signal is a different system. Deciding which roles are mobile-first before the screens are drawn is cheaper than discovering it afterwards.

What the stack decision looks like in 2026

Clients often arrive with a technology opinion. The data suggests it matters less than they think, provided the choice is mainstream. The 2025 Stack Overflow Developer Survey, which had 23,678 respondents answer the web frameworks question, found Node.js used by 48.7%, React by 44.7%, Next.js by 20.8%, Express by 19.9%, ASP.NET Core by 19.7%, Angular by 18.2%, Vue by 17.6%, Django by 12.6% and Laravel by 8.9%. On languages (31,771 respondents) JavaScript led at 66%, with Python at 57.9%, TypeScript at 43.6% and PHP at 18.9%. PostgreSQL (55.6%) was ahead of MySQL (40.5%) among databases.

Developer usage of web frameworks, 2025 Stack Overflow survey (23,678 respondents)
FrameworkShare using it
Node.js48.7%
React44.7%
Next.js20.8%
ASP.NET Core19.7%
Angular18.2%
Vue.js17.6%
Django12.6%
Laravel8.9%
[8]

The practical reading for a Malaysian SME: any of the top eight is hireable locally and will be maintained for a decade. The question to ask a vendor is not "React or Vue?" but "who will maintain this in year three, and will the code and the data be ours?" Pick a stack the business can staff, insist on owning the repository and the database, and make sure the architecture is modular enough that a module can be replaced without a rewrite — BCG’s sixth factor.

A worked example

The decision this exposes is sequencing, not budget. The company that ships the ordering workflow in twelve weeks and then adds purchasing in the next quarter spends about the same over a year as the one that briefs everything at once — and has a working system for nine of those twelve months.

What to decide before the first screen

  • The one workflow the first release must carry end to end in production, and the two or three it explicitly will not.
  • Every role, and for each kind of record whether that role can create, read, edit, approve, export or delete it.
  • Which roles are mobile-first, and what the system must do on a slow connection — what is cached, what queues, what fails loudly.
  • The exceptions: what happens when a record is late, rejected, duplicated, returned or edited after approval.
  • The integrations, named: accounting package, payment gateway, e-invoicing, WhatsApp or SMS, and who owns the credentials.
  • The data volume at eighteen months — rows, files, concurrent users — so pagination, search and reporting are designed for it rather than retrofitted.
  • Ownership: the repository, the database, the domain, the cloud account and the documentation sit in the company’s name from day one.
  • The acceptance test for release one, written before development starts, in the business’s words.

Costs, risks and caveats

Rate data for Malaysia is thin and secondary. Of the fourteen web development agencies visible on Clutch’s Malaysia listing when we checked, nine published an hourly band of US$50–99 and three of US$25–49; six set a minimum project size of US$10,000 and four of US$25,000. Second Talent’s October 2026 rate card — a staffing marketplace’s own observed quotes, not an independent survey — puts Malaysian freelance full-stack developers at US$20–34 an hour for junior, US$32–75 for mid-level and US$45–105 for senior. We quote these as the range the market advertises, not as what a project will cost; a 300-hour build at US$35 and at US$90 are very different invoices for very different levels of ownership and risk.

Two risks are under-priced in most briefs. The first is quality that nobody tests for: WebAIM’s 2026 survey of the top million home pages found detectable WCAG accessibility failures on 95.9% of them, averaging 56.1 errors per page, with low-contrast text (83.9% of pages), missing form labels (51%) and missing alt text (53.1%) the commonest — and a form a staff member cannot operate with a keyboard, or a customer cannot read in sunlight, is a support ticket, not an edge case. The second is performance under real data: the 2024 Web Almanac found only 43% of origins passing Core Web Vitals on mobile, and a web application that is fast with 200 demo rows and slow with 200,000 real ones is the commonest complaint we hear from companies on their second vendor.

What we would do

We would start by writing the workflow and the permission matrix with the people who do the work, not their managers, and we would make the first release deliberately small — one workflow, the mobile roles designed as mobile, the exceptions handled. We would choose a mainstream stack the client can staff, put every asset in the client’s name, and set the acceptance test and the performance budget (the Core Web Vitals thresholds, on a mid-range Android phone, with production-scale data) before a line of code. Then we would ship, measure, and plan the second module from what the first one taught. On the evidence, that is the only approach with better than even odds.

Key takeaways

  • Malaysian demand is not the constraint: 96.8% of households are online and establishments earned RM1.29 trillion from e-commerce in 2024, yet only 58.7% of businesses have their own website.
  • Design for the phone first — 99.4% of Malaysian internet users go online by mobile and only 39.5% by computer — and treat any field or customer role as a mobile role.
  • Small, well-specified projects succeed: Standish found a 71% success rate under US$750,000 in labour against 2% above US$10 million, and PMI traced 47% of missed goals to poor requirements.
  • The framework matters less than ownership and staffability; any mainstream stack from the Stack Overflow top eight is a sound choice if the code, data and accounts are in the company’s name.
  • Set the acceptance test, the permission matrix and the performance budget before development starts, and ship one workflow end to end before briefing the next.

Sources

  1. ICT Use and Access by Individuals and Households Survey Report 2024 (infographic) — Department of Statistics Malaysia, 2025
  2. Internet users by device (ICTHS open data, 2019–2025) — Department of Statistics Malaysia, 2025
  3. Usage of ICT and E-Commerce by Establishment 2025 (reference year 2024) — Department of Statistics Malaysia, 2026
  4. Malaysia’s e-commerce income hits RM918.2 bil in first nine months of 2024 (reporting DOSM Malaysia Digital Economy 2024) — MIDA (secondary, citing DOSM), 2024
  5. CHAOS Summary 2009 — The Standish Group, 2009
  6. Pulse of the Profession: Requirements Management — A Core Competency for Project and Program Success — Project Management Institute, 2014
  7. Flipping the Odds of Digital Transformation Success — Boston Consulting Group, 2020
  8. 2025 Developer Survey — Technology — Stack Overflow, 2025
  9. Top web developers in Malaysia (agency listing) — Clutch, accessed 2026
  10. Cost to hire freelance developers in Malaysia — Second Talent (secondary, vendor rate card), 2026
  11. The WebAIM Million 2026 — WebAIM, 2026
  12. Web Almanac 2024 — Performance — HTTP Archive, 2024

Choose how to get started

Build with us

A system engineered around how you work.

Scoped in a proposal, designed as a prototype, delivered in milestones you can review.

Start a project

See the work

Systems already running in production.

Marketplaces, operations portals and mobile apps, with the case studies behind them.

View our work
WhatsApp