Business · Investment

Three in four Malaysian SMEs have gone digital. Just over half call it a success. How to buy software so it works

Owners of traditional businesses are right to be cautious: digital projects fail often, and for reasons surveys keep repeating. Seven common objections to investing in an internal system, answered honestly, and a way to buy in stages so that no single decision can sink you.

Pikabyte Engineering 6 min read

43%

of small enterprises in Malaysia reported success with digitalisation, against 67% that had adopted it, in UOB's Business Outlook Study 2026

In UOB’s Business Outlook Study 2026, which interviewed 268 small and medium enterprises in Malaysia, 75% had adopted digital tools and 54% reported success with them. Among the smaller firms the figures were 67% and 43%.

The study also asked what held the efforts back. The most common answers were that digitalisation was expensive to implement, at 37%, and that employees lacked digital skills, at 36%. Concerns about cybersecurity and running out of funds to continue followed.

An owner who has built a business over thirty years without a computer system is not being backward by hesitating. The evidence says a good share of this money is spent badly. This article takes the usual objections one at a time.

First, why projects fail

The large studies are about large companies, so read them as a warning about scale.

Boston Consulting Group reported in 2020, from work with 70 companies and a survey of 825 executives, that 70% of digital transformations fall short of their objectives. A Gartner analyst said in November 2025 that 70% of ERP initiatives fail to fully meet their original goals and 25% fail catastrophically. The reasons given were weak business cases, overambitious scope, heavy customisation and too little attention to how people would change the way they work.

Each of those reasons is a symptom of buying too much at once. A small business has an advantage here. It can buy one small thing, see it work and then decide.

“It costs too much”

It may. Some processes are not worth automating, and we say so in our costing guide.

Three points make the sum less frightening.

  • You are pricing one job, not a transformation. Most of our custom projects fall between RM30,000 and RM150,000, and a first project covers a single process.
  • The saving can be counted in advance. Hours a week, mistakes a month, days of delay. In our experience most projects pay back within six to eighteen months.
  • It can be financed. Bank Negara Malaysia’s SME Automation and Digitalisation Facility lends up to RM3 million for software and IT services at up to 4% a year. Budget 2027 extended the accelerated capital allowance on ICT equipment and software to the end of 2030. See who pays for your AI project.

The cost of doing nothing also rises. The minimum wage goes from RM1,700 to RM2,000 in June 2027, with firms below RM50 million in annual sales exempted for now.

“We tried a system before and nobody used it”

This is a common objection, and it is usually true.

What happened, as a rule, is that the company bought a general package and asked staff to change how they work to fit it. They kept the spreadsheet on the side, and within a year the spreadsheet was the real system again.

An internal tool is built the other way round, from the way your best clerk already does the job. Two habits make the difference.

  • The people who do the work test it every week, from the second week.
  • It is designed for where the work happens. The quality control system we built for a manufacturer uses phone camera scanning made for gloves, glare and a factory floor, with manual entry as a fallback.

“My staff are not computer people”

The UOB figure agrees with you: skills were the second most common obstacle.

The answer is to ask less of them. One screen per job. Large buttons. A phone, which every member of staff already uses all day, in place of a desktop. Words they use, in the language they use. If a new hire cannot do the job on the system after an hour, the system is wrong.

If you pay the HRD Corp levy, training can be claimed against it, as we explain in our funding guide.

“I will be at the mercy of the vendor”

You will be, if you sign the wrong contract. Ask for four things in writing.

  1. You own the source code, and receive it.
  2. The system runs in a cloud account in your company’s name, which you can lock a vendor out of.
  3. Your data can be exported in a standard format at any time.
  4. The system is built with common, well-documented technology that another competent team could take over.

A vendor who resists any of these is telling you something. Our checklist for choosing a software partner has the rest.

“Our business is different”

It is, and that is the argument for building. If your process were ordinary, a package off the shelf would be cheaper and you should buy it. The parts worth building are the ones that are peculiar to you: how you price, how you inspect, how you look after your best customers. We set out the test in build or buy.

“What about hackers and our data”

Cybersecurity concerns were among the top obstacles in the UOB study. It is a fair worry, and it should be compared with the present arrangement honestly. Today the customer list is in a spreadsheet on a personal laptop, the price list is in a WhatsApp group that includes people who left last year, and nobody can say who changed what.

A properly built system has individual logins, permissions by role, a record of every change and backups. For a small team the essentials take about a day to put in place, as we describe in the one-day AWS security baseline.

“AI is changing everything. Better to wait”

Waiting has a cost that is easy to miss. An AI assistant is useful to a business only when it can read that business’s records and documents. Every month your orders stay in a chat group is a month of history no future tool will be able to use.

The system you build now is what makes AI possible later. Once quotations, orders and procedures are in one place, an assistant that answers from them is a modest addition.

How to buy in stages

You do not have to commit to the whole thing at once. Break the decision up.

Stage 1: a short paid study. One or two weeks. The vendor watches the work being done and writes down the process, the rules, the volumes and a fixed price for building one job. You own that document and can take it to anyone.

Stage 2: one job, at a fixed price, live within weeks. A first system does not need a year. The quality control system for Soon Chuan Hong Hardware went from kickoff to production in 9.5 weeks. A litigation platform we built for a law firm took six, and cut the time its staff spent filing and checking by 95%.

Stage 3: measure against the numbers you counted. Hours, mistakes, delay. If it did not pay, stop here. You have lost a bounded amount and learned a great deal about your own process.

Stage 4: the next job, chosen from what the first one showed you.

At every stage you can walk away with something you own.

Where the successful firms differ

In the same UOB study, 43% of respondents said digitalisation had increased their productivity and efficiency. The Malaysia Productivity Corporation reports that 59.7% of firms saw higher productivity after digital adoption. Both figures are what firms say about themselves, and neither is close to 100%.

The Malaysia Productivity Corporation’s explanation of the difference is the most useful sentence in its report: “Digitalisation lifts productivity when it is matched by better workflow design, managerial capability and workforce skills”.

Software alone does not do it. A clear process, an owner who uses the figures, and staff who were involved from the start do. All three are in your hands, and none of them costs much. For what the first project usually looks like, see what an internal tool replaces first.

Sources

Written by

Pikabyte Engineering

Pikabyte builds custom business systems, applied AI and secure cloud infrastructure from Malaysia. Talk to us.

Start a conversation

Want this applied to your business?

Tell us what is slowing the business down. You will talk to an engineer, not a salesperson, and get a straight answer on whether software is the right fix.