Content localisation in Asia is not translation: three market lessons

Content localisation in Asia means rebuilding a message around each market's own competitors, audience tensions and trust signals, not translating a global script. SOMIN's work in Hong Kong, the Philippines and Singapore shows the gap: Keeta's strongest lines dropped price talk, P&G's category addressed different tensions in search and on Meta, and Singapore mothers trusted peers more than brands.

Key takeaways

• Language is necessary but not sufficient: the argument, the proof and the messenger also change by market.
• In Hong Kong, an audience check turned Keeta's price-led lines into curation-led lines, written in English and Cantonese.
• In the Philippines, localisation also meant aligning channels and reading a crowded local competitive set before choosing a lane.
• In Singapore, Mothercare's local read showed that peer and creator endorsement mattered more than brand announcements.
• A localisation brief should start from local competitor and audience data, and only then move to copy.

Why is translation not enough in Asian markets?

Translation is not enough because most Asian markets are multilingual and the deciding differences sit in what people value, not only in which words they read. Getting the language right is the entry ticket, not the message.

Look at two cities that both use plenty of English. In Hong Kong, 96.0% of people aged five and over could speak Cantonese in 2021, according to the 2021 Population Census, while 57.7% could speak English and 56.5% Putonghua. In Singapore, the Census of Population 2020 found English was the language most frequently spoken at home for 48.3% of residents aged five and over, up from 32.3% in 2010, with Mandarin at 29.9%.

Those mixes shape the script, but in each case below the language was handled and the message still had to change. That is the difference between translation and localisation.

What did Hong Kong teach us about price and quality?

Hong Kong taught us that the category's loudest message can be the wrong one for your audience. Keeta entered a food-delivery conversation built on price and convenience, and the data said its audience wanted something else.

SOMIN benchmarked 145 pieces of paid social and search media from Keeta HK, FoodPanda and Deliveroo, from 22 January to 22 March 2025. The dominant tension was the need for affordable, convenient food, especially in festive periods and for families and friends. FoodPanda dominated it, Deliveroo also addressed it, and Keeta lagged behind both rivals on it.

A GWI snapshot of Hong Kong food orderers in the above-average income bracket then changed the brief. This audience under-indexed on budget-conscious, value-driven behaviour and over-indexed on curated, quality-first choices. So every direct-to-action line was rewritten.

"Quick to your door, kind to your wallet" became "Smart meals for smart people—Keeta delivers fast, curated, and made for you." Across four stories and three tones, that produced 36 raw lines and 36 refined ones in English and Cantonese. The Keeta HK case study is a pre-launch creative sample, not a report of campaign results, but the lesson holds: translating the category's price message into Cantonese would have missed the audience entirely.

What did the Philippines teach us about channels and competition?

The Philippines taught us that one market can need different messages on different channels, and that the local competitive set decides which lanes are open. Localisation happens inside a market as well as between markets.

For P&G, SOMIN compared Google Search text ads and Meta content from May to July 2024. Unilever accounted for 80.43% of the search text ads, led by its cleanipedia (208) and Allthingshair (155) sites, while P&G's biggest search presence came from Head & Shoulders (45) and Pantene (41). On Meta, the radiant-skin tension drew 1,955,781 engagements from 22 pieces of media, against 399,517 from 39 pieces on skin concerns and confidence. The tensions brands addressed on Meta were not always the ones they addressed in search, so the P&G Philippines case study recommended aligning the two, with search-moment keyword lists for each tension.

Tonik Bank faced a different problem: a crowded field. SOMIN read 126 Tonik Google ads against 516 from seven Philippine digital banks. Tonik's share peaked at 54% of the sampled window from December 2025 to February 2026 and then went silent for about five months, while MariBank came to run 70% of the latest window. All nine tensions already had brand or competitor presence, so the Tonik Bank case study advised differentiating inside contested lanes and delivered 88 banners with every message at nine words or fewer.

The people making content are part of the context too. In SOMIN's creative-content workspace, which tracks AI content tools, a theme on marketing and operational tensions in the Philippines grouped 57 posts around manual admin, pressure to deliver quality content at scale without losing authenticity, hustle culture and burnout. A localisation process that adds work for stretched local teams will not survive contact with them.

The calendar is local as well. Philippine state broadcaster PTV News marks 1 September as the start of the "ber" months and the country's long Christmas season, months before most markets think about festive content. Our APAC festive marketing calendar covers those timing differences.

What did Singapore teach us about trust?

Singapore taught us that the messenger can matter as much as the message. Mothercare Singapore had clear direction from global HQ, but its local team was tracking more than 4,000 category ads a year by hand and still guessing what would land with Singaporean parents.

SOMIN's competitor read grouped the category's content into themes and compared engagement per post. Parenthood-challenge content (500 pieces of media, 19,302 engagements) yielded about 39 engagements per post, against about 23 for affordable, high-quality product content (1,120 pieces, 25,505 engagements). GWI data then showed that Singapore mothers over-index on endorsements by celebrities or influencers (index 147.3) and on recommendations or comments on social media (index 131.0).

The tone brief followed from the data: fun, bold, modern, stylish and slightly cheeky, built with young, relatable creators instead of authority-figure spokespeople. The Mothercare Singapore case study reports 80% less time on audience research and idea development, 74% cost savings against traditional execution and more than 25 hours a week reclaimed by the team.

How does localisation differ from translation in practice?

Translation keeps the argument and changes the words. Localisation checks the argument against local data first.

What should a localisation brief for Asia include?

A good brief answers six local questions before it asks for copy. Each answer should point to data from that market, not from the lead market.

1. Who are the local competitors, and which tension does each one own?
2. What does the local audience over-index and under-index on?
3. Which channels carry which tensions, and do they match?
4. Who is the trusted messenger: the brand, an expert, peers or creators?
5. Which languages and scripts does each channel need?
6. Which local dates and seasons change the timing?

SOMIN's strategy agent answers the first four from live competitor and audience data and writes the brief in one run, which keeps the extra work off local teams. The last two still need a local check: script choices and holiday dates should come from people and official sources in the market.

FAQ

Take one regional campaign and run the six questions above for a single market this month. If the answers match your global brief, translate with confidence. If they do not, you have found the part of the message that needs rebuilding.