Skip to content
    Get in touch

    Sources

    Where our numbers come from

    Every outside figure on this site is listed here with the document it came from. If an outside figure we quote is not on this page, we should not be quoting it.

    What gets in

    A figure is listed only if a named organisation stands behind it, the method is disclosed, the document can be retrieved, and we attribute it wherever it appears. Company and consultancy research is held to the same test as academic work, and passes it as often. What fails is a number with no sample, no field dates and no document behind it, however many places repeat it.

    Two dates, not one

    Each entry shows when it was published and the period the data covers. Those are different things, and a recent report can carry old data. Claims about technology and market adoption are held to 2024 or later. Findings about how people behave are allowed to be older, and carry their year in plain sight so you can judge them yourself.

    Malaysian market and business conditions

    Regulation

    Messaging and chatbots

    • 72.4% of people surveyed said they are more likely to buy from a business they can message.

      Meta with Kantar. State of Business Messaging. n=11,056 across 22 markets, fielded April-September 2025

      Publication date not stated · Data covers Fielded April to September 2025

      Malaysia is in the sample but the published figures are global aggregates, not a Malaysian breakdown. Meta has a direct commercial interest in business messaging.

      https://business.whatsapp.com/resources/resource-library/state-of-business-messaging

    • AI chatbots resolve 44.8% of chats without escalating to a person.

      Comm100. 2026 AI Live Chat Benchmark Report (figures as published in the linked article)

      Published 2026-03-12 · Data covers 2025 for the chatbot CSAT figure; the resolution-rate period is not stated

      Vendor benchmark data.

      https://www.comm100.com/blog/what-percentage-of-chats-can-ai-chatbots-resolve/

    • Sharing over private messaging in Malaysia rose from 29.4% to 48.6% in two years.

      Malaysian Communications and Multimedia Commission (MCMC). Internet Users Survey 2024. Figure 28, "Online content sharing platform", printed page 39 (PDF page 36)

      Published 2025-10-06 · Data covers Fieldwork 1 October to 31 December 2024

      This measures which platforms people share CONTENT on, not how customers contact businesses.

      https://www.mcmc.gov.my/skmmgovmy/media/General/pdf2/Internet-Users-Survey-2024-Report.pdf

    Appointments and reminders

    • At Hospital Sultan Ismail, Johor Bahru, appointment no-shows fell from 40.7% to 22.2% after an automated SMS reminder was introduced.

      Ministry of Health Malaysia authors, Hospital Sultan Ismail Johor Bahru. Jurnal Sains Kesihatan Malaysia 19(01)

      Published 2021 · Age disclosed on every page that uses it: 2021

      A single-site pre/post quality-improvement study, not a randomised controlled trial.

      DOI 10.17576/jskm-2021-1901-01

    • 28% of outpatient appointments at Hospital Kuala Lumpur ended in a no-show, across 246,943 appointment records.

      Abdullah Fahim Ahmad Hamdan and Azuraliza Abu Bakar. Machine Learning Predictions on Outpatient No-Show Appointments in a Malaysia Major Tertiary Hospital. 246,943 records from HKL’s Patient Management System, 2019 data; 69,173 no-shows

      Published 2023-10-30 · Age disclosed on every page that uses it: 2023, using 2019 data

      Hospital Kuala Lumpur is a large public tertiary referral hospital. It is not safe to present 28% as what a private GP, dental or aesthetic clinic experiences.

      https://pmc.ncbi.nlm.nih.gov/articles/PMC10624443/

    • Pooled across randomised trials, SMS reminders do not reach statistical significance; phone calls do.

      Journal of Hospital Management and Health Policy. Systematic review of appointment reminder interventions

      Publication date not stated · Age disclosed on every page that uses it: pooled review, undated — cite as a body of trial evidence, never as a dated figure

      https://jhmhp.amegroups.org/article/view/10215/html

    Website value and AI-assisted discovery

    • 83% of US small businesses now have a website, up from 64% in 2018; the 17% still without one mostly cite reasons other than cost.

      Clutch. Small Business Websites in 2025: Growth, Gaps, and Where to Invest Next. SurveyMonkey panel, n=406 US small-business owners, quota-balanced by age (18-99)

      Published 2025-08-21 · Data covers Fielded 5 August 2025

      US small businesses, self-reported online panel. Clutch is a B2B marketplace connecting businesses with web-design agencies — a direct commercial interest in "you need a website."

      https://clutch.co/resources/state-of-small-business-websites-2025

    • 82% of US small employers have their own website; adoption rises sharply with headcount, from 75% at 1-9 employees to 97% at 50+.

      National Federation of Independent Business (NFIB). 2025 NFIB Technology Survey (press release: "Latest NFIB Survey Reveals How Main Street Business Owners Adopt New Technology"). Emailed to a random sample of 20,000 NFIB members; 521 respondents completed it (~2.6% response), spanning services (21%), retail (16%), construction (16%), manufacturing (14%)

      Published 2025-06-25 · Data covers Fielded 6-31 March 2025

      NFIB-member sample: skews toward established, employer small businesses, not the broader self-employed/non-employer population.

      https://www.nfib.com/wp-content/uploads/2025/06/2025-NFIB-Technology-Survey.pdf

    • Consumers judge a website’s credibility in under 20 seconds; a majority have immediately left a site that looked untrustworthy, and slow load time or typos are named trust-breakers.

      Clutch. What Trustworthy Websites Look Like in 2025 (According to Consumers). SurveyMonkey panel, n=263 US consumers who are household purchase decision-makers, quota-balanced

      Published 2025-07-22 · Data covers Fielded 24 June 2025 (page last updated 2026-03-24)

      Small US-only sample; self-reported perception, not measured behaviour. Clutch sells web-design leads.

      https://clutch.co/resources/what-trustworthy-websites-look-like-2025

    • 97% of consumers read reviews before choosing a local business, and the share who read them every time rose to 41%, up from 29% a year earlier.

      BrightLocal. Local Consumer Review Survey 2026: Star Ratings Keep Rising, Old Reviews Don’t Cut It. SurveyMonkey representative panel, n=1,002 US adult consumers (age-split 18-29 22%, 30-44 25%, 45-59 28%, 60+ 25%); no fieldwork window disclosed on the page beyond the publication date

      Published 2026-02-11

      US-only, single vendor. No fieldwork window disclosed, only a publish date.

      https://www.brightlocal.com/research/local-consumer-review-survey/

    • Consumers using AI tools to get local-business recommendations jumped from 6% in 2025 to 45% in 2026 — AI is now the third most-used channel for this, behind only Google and Facebook. ChatGPT leads AI tools specifically (used by 31% in the past 12 months), then Google AI Mode (23%) and Gemini.

      BrightLocal. Half of Consumers Are Asking AI for Business Recommendations (Local Consumer Review Survey 2026 AI mini-report). SurveyMonkey representative panel, n=1,002 US adult consumers; the AI-specific figures are drawn from the 455 respondents who said they had used an AI tool for local-business recommendations in the past 12 months. No fieldwork window disclosed.

      Published 2026-03-10 · Age disclosed on every page that uses it: 2026, measuring a rise from 2025

      US-only, single vendor, fast-moving figure.

      https://www.brightlocal.com/research/lcrs-ai-trust/

    • 73% of a recent local-business search started on mobile (19% computer, 8% tablet), but only 8% of searchers chose an AI tool as their very first channel — and of those, only 18% felt ready to contact the recommended business without checking further.

      BrightLocal. Where Are Your Customers Really Searching? (Consumer Search Behavior 2026, Part 2: Channels). SurveyMonkey representative panel, n=1,227 US consumers who had searched for a local business in the past 3 months. No fieldwork window disclosed beyond the 3-month recency screen and publish date.

      Published 2026-07-09

      US-only.

      https://www.brightlocal.com/research/consumer-search-behavior-channels/

    • AI search tools (ChatGPT, Gemini, Perplexity, etc.) accounted for less than 1% of referral traffic across the sites studied in 2025, and organic search still converted visitors better than AI-search referrals in every industry measured.

      BrightEdge. AI Search Visits Surging in 2025 — But Organic Search Remains the Cornerstone of Digital Growth. Described as "thousands of queries and top-performing websites, including many Fortune 100 brands" — an exact sample count is not given.

      Published 2025-09-12 · Data covers January-August 2025

      Sample skews toward large/enterprise, top-performing sites, not small businesses; exact sample size never stated as a number. BrightEdge sells SEO software with a direct interest in an "organic search still wins" narrative.

      https://www.brightedge.com/news/press-releases/brightedge-data-finds-ai-accounts-less-1-search-organic-traffic-continues

    • 76% of online shoppers globally prefer to buy where product information is presented in their own language, and 40% say they will never buy from a website in another language.

      CSA Research, with Kantar. Can’t Read, Won’t Buy — B2C. n=8,709 verified consumers, screened from 31,933 respondents, across 29 countries

      Published 2020-07-07 · Age disclosed on every page that uses it: 2020

      Global figure, six years old. Malaysia is not confirmed among the 29 countries sampled (Indonesia is the nearest named market).

      https://csa-research.com/l/media/Consumers-Prefer-their-Own-Language

    • In Malaysia, desktop still edges out mobile for general web page views: 64.77% desktop vs 35.23% mobile in August 2026.

      StatCounter GlobalStats. Desktop vs Mobile vs Tablet Market Share Malaysia. Aggregated page views across StatCounter’s network of monitored sites (StatCounter states over 5 billion page views/month across 2+ million websites globally); the Malaysia-specific site count behind this chart is not disclosed.

      Published live dataset, queried 2026-09-07 · Data covers August 2026

      Self-selected network of sites carrying StatCounter’s tracking script, not a representative population panel, and it runs counter to the popular "Malaysia is mobile-first" narrative. Not specific to business-website traffic.

      https://gs.statcounter.com/platform-market-share/desktop-mobile/malaysia

    What we deliberately do not claim

    Some numbers we would like to quote do not exist, and some that circulate widely fall apart when traced. Both are listed here, because leaving them out silently would be the easier and worse choice.

    How many salons, clinics or cleaners there are in Malaysia

    No count exists. The Department of Statistics Malaysia’s open-data catalogue carries no establishment-count dataset by industry class at any MSIC level — every endpoint returns a 404. We therefore put no market-size figure on any trade page.

    The share of Malaysians who use WhatsApp

    A single figure circulates across dozens of marketing pages, each crediting the same two annual reports. We fetched both. Neither contains any WhatsApp statistic at all. We then read the regulator’s own 2024 survey in full: it does not carry one either. Three independent checks, three times nothing, so we do not repeat it.

    The e-Invoice turnover threshold

    The tax authority’s own FAQ and its own guideline give different figures for the turnover below which a business is exempt. Until that is settled, we publish the phase dates, which are not in dispute, and no threshold.

    Results we have achieved for customers

    We do not publish outcome numbers, client counts or testimonials, because we have not earned them yet. What we describe instead is what the system does, which you can watch it do.

    Found a problem

    If a figure here is out of date, misread, or no longer says what we say it says, tell us and we will correct it. That includes the people we compete with.