Localisation is not translation

The engineering team exports the string file, a Bahasa Indonesia translator works through it in five days, the strings are pushed to production, and the product is declared localised. What has actually happened is that the words changed. The product's behaviour — its trust architecture, its information hierarchy, its social proof, its payment assumptions — remained entirely Western.

What gets translated vs. what needs to change.

Consider what a ride-hailing app has to do to enter Indonesia from Singapore properly. Translating "Book a Ride" to "Pesan Sekarang" is the easy part. The real work is rethinking things like when the driver's photo and plate number appear — users may need to see them before booking, not after, because the trust calculation is different — and whether a cancellation-penalty flow designed for Singapore creates friction and social media backlash in Jakarta, where users may cancel more often and for different reasons.

Large platforms have the resources to do this properly. Most founders expanding to Indonesia, Thailand, or Vietnam do not do this. They translate, ship, and then wonder why their activation and retention metrics are far lower than they were at home. The answer is usually visible in 30 minutes of watching Indonesian users try to navigate the product without any guidance — which you cannot see if you never sit with Indonesian users in their actual environment.

Real localisation has four components that are not translation: trust architecture, information hierarchy, social proof relevance, and payment behaviour design. Each takes longer than the translation itself. Each requires actual user research in-country, not remote sessions through a Zoom call with a translation assistant. And each will surface product decisions that your engineering team made for a Western user context that do not survive contact with a Southeast Asian one.

Trust architecture: what signals credibility in this market.

Trust signals are market-specific. In the US, a SaaS product builds trust through clean design, published security certifications (SOC 2, ISO 27001), G2 review badges, and logos of recognisable customer brands. In Indonesia, these signals either do not register or actively create uncertainty.

The trust signals that work in the Indonesian market are a different set entirely. Bank transfer payment options signal stability — a company that accepts BCA Virtual Account or Mandiri transfer is implying it has a relationship with an Indonesian bank, which implies institutional legitimacy. Registration as an electronic system operator (PSE) with the communications ministry (Kominfo, now Komdigi), while technically a compliance requirement, is increasingly used as a trust signal by Indonesian B2B buyers because it indicates regulatory seriousness. Indonesian customer logos — not US enterprise logos, not even Singapore company logos — are the social proof that converts. An endorsement from a well-known Indonesian entrepreneur visible on Instagram carries more weight with an Indonesian SME buyer than a Harvard Business Review case study.

For consumer products, the trust architecture shifts again. Tokopedia and Shopee have trained Indonesian consumers to look for "resmi" (official) store badges, seller ratings backed by a large number of reviews, and product images that show the item in Indonesian life contexts — a skincare product photographed in a Jabodetabek apartment, not a sterile studio. Fashion products with models who look Southeast Asian, not European, convert at significantly higher rates. These details seem obvious once stated, but the majority of Western product expansions into Indonesia fail to implement them consistently.

Information hierarchy: not all users read the same way.

Western product design is heavily influenced by the F-pattern reading model — users scan left to right, top to bottom, with attention concentrated in the upper left. This is a real phenomenon for left-to-right language readers. It does not reliably describe how Indonesian and Thai users process product interfaces, particularly mobile interfaces.

A common observation from mobile UX work in Jakarta and Bangkok is that Southeast Asian users give more attention weight to bottom-of-screen elements than Western users, partly because of thumb-zone patterns on larger smartphones common in the region (Samsung Galaxy S and A series, Xiaomi Redmi, Oppo), and partly because of different visual scanning habits developed through consumption of content on platforms like TikTok and Instagram Reels where the action triggers are at the bottom.

More importantly: the information Indonesian users want at the start of a product experience is not the same as what American users want. American users want to understand what the product does (features). Indonesian users want to understand who else is using it and whether it is trustworthy (social validation), then what it costs (price transparency), and only then what it does (features). A Western onboarding flow that leads with feature demonstration before establishing social proof will lose Indonesian users at step one — not because the feature is irrelevant, but because the trust decision has not been made yet.

The practical test: conduct a 5-second test with 10 Indonesian users on your current homepage or app store listing. Ask them what they understood and whether they would download or sign up. Do the same in English with 10 US or UK users. The gap in comprehension and intent will tell you more about your localisation gap than any analytics dashboard.

Social proof that is actually relevant.

The social proof problem is one of the most consistently underestimated in cross-border product expansion. Founders paste their US testimonials into the Indonesian product page — a quote from a VP at Shopify, a logo from a Series B edtech company in San Francisco — and expect it to function as validation. It does not. Indonesian buyers have no frame of reference for these companies. They do not know if they are prestigious or ordinary, large or small, recently successful or recently bankrupt.

Relevant social proof for the Indonesian market means: testimonials from Indonesian business owners with visible Instagram or LinkedIn profiles, preferably in industries the target buyer recognises. Case studies set in Indonesian cities, with Indonesian context (rupiah revenue figures, local market references, Indonesian regulatory challenges). Video testimonials are significantly more trusted than text testimonials in the Indonesian market — a 60-second vertical video of an Indonesian business owner talking about their experience with your product, posted to Instagram Reels, will drive more conversion than a polished written case study on your website.

This social proof does not exist when you first enter the market. Building it requires giving your first 10-20 Indonesian customers exceptional service, then proactively producing content with them — not extracting testimonials through a survey form. Someone from your team, ideally with video production skills, should spend a day at each customer's premises in the first year, filming content that serves both your marketing needs and the customer's own social media needs. Indonesians are enthusiastic content creators; the documentary of their business success is something they will share widely if it makes them look good.

Payment behaviour: not just methods, but psychology.

Payment localisation is frequently treated as a matter of adding GoPay and OVO to the checkout. This is necessary but not sufficient. The psychology around payment in the Indonesian and broader SEA consumer market differs from the West in ways that affect your pricing architecture, your trial model, and your subscription flow design.

Indonesian consumers have a strong preference for prepaid/top-up payment models over subscription models. This is partly because many Indonesians manage household finances on a weekly or semi-monthly cycle rather than a monthly one (salary disbursement patterns vary widely). It is also because subscription charges that hit unexpectedly during a low-cash week create negative emotional responses that drive cancellation and negative reviews — the "I forgot I subscribed to this" anger is real and it drives higher-than-expected churn.

The design implication: build a credit/top-up model alongside your subscription tier. Give users the ability to load IDR 100,000 (approximately $6) and consume the product until it runs out, then reload. This is not the highest-revenue model — subscriptions monetise better long-term — but it is the model that onboards users who are not yet ready to commit to a recurring charge, and it often converts to subscription after 2-3 credit cycles when trust is established.

Thai and Vietnamese markets have similar dynamics with some variation. Thai consumers are more comfortable with credit cards than Indonesians but show high sensitivity to auto-renewal. Explicit consent flows for auto-renewal, with clear cancellation instructions, significantly reduce chargeback rates and negative reviews in Thailand. In Vietnam, MoMo wallet integration is non-negotiable for consumer products, and the cashback and reward mechanic that MoMo promotes needs to be incorporated into your pricing strategy — Vietnamese consumers comparison-shop aggressively through cashback lens.

• 01
  Conduct in-person user research before engineering localisation. Hire a UX research firm in Jakarta, Bangkok, or Ho Chi Minh City to run 8-12 contextual inquiry sessions with target users before your engineers touch the string file. What you learn will change the product requirements significantly, and changing them before the translation sprint is far cheaper than rebuilding after. Remote platforms such as Dscout and local research firms such as Populix (Indonesia) can run these sessions affordably.
• 02
  Build local trust signals into the product architecture, not the marketing layer. Indonesian bank transfer options, PSE registration details, local customer logos, and Indonesian support contact details need to be in the product core — checkout flow, pricing page, app store listing — not just on the marketing website. If a user sees them on the landing page and then does not see them during checkout, the trust signal is lost precisely when the purchase decision is being made.
• 03
  Produce original social proof, do not repurpose Western testimonials. Allocate budget for a content producer to spend one week in-country per quarter, producing video testimonials, case study content, and behind-the-scenes footage with your best Indonesian or Thai customers. This content doubles as the customer's own social media content — they will share it — and your conversion asset. The cost is modest relative to paid acquisition, and the asset keeps converting long after a paid campaign stops.
• 04
  Design payment flows for the prepaid psychology, then upsell to subscription. Offer a credit top-up option for first-time users who are not ready to subscribe. Set a clear trigger for the subscription upsell — for example, when a user's credit balance drops below 20%, send an in-app message explaining that a monthly subscription saves them 30% and auto-refills. This motion respects the financial caution of the market while building toward the subscription revenue model you want at scale.
• 05
  Test with native speakers who are not your translators. Your translator will tell you the Bahasa Indonesia is correct. That is their job. You also need 5-10 native speakers who have never seen your product to complete specific tasks and tell you where they get confused or suspicious. These are different people doing different things. The translator ensures linguistic accuracy; the testers ensure behavioural fit. Budget for both. The latter is cheaper than a re-design sprint after a failed launch.
• 06
  Localise your error states and empty states, not just your main flows. Nothing breaks trust faster than an error message in English or a technically-correct but culturally-awkward Bahasa Indonesia error message when something goes wrong. "Terjadi kesalahan" (an error occurred) is technically correct and emotionally cold. "Maaf, ada gangguan sebentar — tim kami sedang mengatasinya" (Sorry, there is a brief disruption — our team is working on it) is warmer and more consistent with Indonesian business communication norms. Error states are read at moments of maximum frustration; they should be written by a native speaker with customer experience instincts, not a translator.

Frequently asked questions.