In the United States and most of Western Europe, the app store is the first place a user looks for something new. The funnel runs from ad to store listing to install, and UA teams build around that path. In Korea, Japan, and much of Southeast Asia, the messaging app sits in that first position instead. KakaoTalk, LINE, Zalo, and WeChat are the screens people open dozens of times a day, and that frequency turns them into discovery surfaces rather than just chat tools.
When one app reaches almost everyone in a market, app discovery routes through it by default, not by choice.
Three habits make these apps the home screen rather than one icon among many:
- Bill pay, transit, and identity live inside the app. WeChat handles payments, government services, and QR-based checkout. KakaoTalk runs KakaoPay and KakaoT for rides. Users keep the app open because daily errands depend on it.
- Sharing happens in chat, not in a browser. A friend who finds a game sends it as a message, a gift, or a mini-program link. The recommendation arrives in the same surface where the user already spends time, so the path from interest to install is short.
- The platform hosts its own apps. Mini-programs and mini-apps run inside the messaging shell. A user can play a game or order food without ever leaving the app or touching the store.
This matters for UA planning because the Western model assumes a store visit you can measure and a creative flow you control. Inside a messaging ecosystem, a meaningful share of installs starts from a shared link, a gift, or an in-app ad placement, and the platform owns the rules for each. Your creative competes against chat content, not against other store listings. Your targeting depends on signals the platform exposes, which differ from what Meta or Google offer.
The practical effect is that a campaign tuned for App Store Search Ads or Google UAC will not map cleanly onto these markets. The placements, the creative norms, and the attribution paths all change. A budget that ignores the messaging layer in Korea or Vietnam leaves the main distribution channel untouched and competes only for the fraction of users who still discover apps through the store.
The sections that follow break down how each platform turns chat into installs, where the mechanics differ by country, and what to adjust in measurement and budget before you commit spend.
How KakaoTalk turns chat into an install channel
KakaoTalk is not a single ad placement with one set of rules. It is a stack of surfaces, and an install can reach your app through any of them. Treating it as “the Korea version of Meta” misreads the structure. The platform routes installs through 4 distinct paths, each with its own targeting, creative format, and cost profile.
Biz board: the ad slot inside the chat list
The most direct path is the Biz Board placement that sits at the top of the chat tab, above the conversation list. Because users open the chat tab dozens of times a day, this slot earns repeat impressions from the same person without them leaving their main screen. The format is compact: a small image, a short line of copy, and a tap target that can open a landing page, a KakaoTalk channel, or a deep link into your app.
For UA, the constraint is space. You have a single image and a few words to earn the tap, so the creative job is closer to a push notification than a video ad. Teams that port a 15-second App Store creative into this slot tend to underperform, because the format rewards a clear offer stated in one line over a produced spot.
Kakao gifting: install through a sent item
KakaoTalk’s gifting feature lets users send products, vouchers, and digital goods to a contact inside a chat. For app marketers, this turns an install into something a friend hands you rather than something you search for. A user can receive an in-game item, a coupon, or a subscription credit as a gift, and redeeming it pulls them into your app.
The mechanism matters because it changes who initiates the install. In a store-search flow, the user already wants the app. In a gift flow, the sender wants the recipient to have it, which front-loads intent on the wrong person. That shifts your creative target: you are persuading the gift-giver, not the end user, and the message that sells a gift differs from the message that sells a download.
Mini-apps: install without a store visit
KakaoTalk hosts mini-apps that run inside the messaging shell. A user can open a service, play, or transact without going to the App Store or Google Play. For some categories, the mini-app is the product surface, and a separate native install is a later step rather than the first one.
This breaks the assumption that an install is the start of the relationship. A user might use your mini-app for weeks before a prompt nudges them toward the full native app. If your measurement counts only the store install, you miss the engagement that happened earlier inside Kakao, and you misread where the conversion really started. Plan for a 2-stage funnel: mini-app trial first, native install second, with separate signals for each.
The games tab: a category storefront inside chat
KakaoTalk runs a games tab that functions as a curated storefront for mobile titles, with featured slots, rankings, and social hooks that show which of your contacts already play. The social proof here is built in. A user sees that 3 friends play a title before they decide, and that friend signal does work that a paid creative cannot.
For games UA in Korea, ignoring this tab means competing only on paid placements while organic discovery flows through a channel you never touched. The trade is that placement and featuring depend on the platform’s curation, so you cannot simply buy your way to the top slot the way you would bid on a keyword.
What the 4 paths mean for planning
The reason to separate these surfaces is that they fail and succeed for different reasons, and a single campaign number hides which one is working.
- Biz Board rewards tight, offer-led creative and repeat exposure. Measure it on cost per tap and post-tap install rate.
- Gifting depends on the giver’s intent, so track redemption rate and the gap between gift sent and app opened.
- Mini-apps produce engagement before the native install, so you need an event upstream of the store to read true intent.
- Games tab trades paid control for social distribution, so judge it on assisted installs and friend-graph reach, not last-click alone.
A team that buys only Biz Board impressions and calls that “running KakaoTalk” has touched 1 of 4 channels. The platform’s value comes from the fact that the same 90%+ population reach can be approached through an ad slot, a gift, an in-app service, and a social storefront at once, and each routes a different kind of user to your app. Mapping your product to the path that fits it, a game to the games tab and gifting, a utility to mini-apps and Biz Board, decides whether the spend lands or competes in the wrong format.
LINE, Zalo, and WeChat: the same pattern, different mechanics
The KakaoTalk model repeats across the region: a messaging app builds an ad system on top of chat, adds a mini-program layer, and routes installs through both. The pattern holds. The mechanics do not transfer. A creative and targeting setup that works on LINE will break the rules on WeChat, and Zalo runs on a different signal set than either. Plan per platform, not per region.
LINE in Japan
LINE runs ads through LINE Ads (also called LINE Ads Platform), and the inventory spans the chat list, the LINE VOOM feed, LINE News, and the wallet tab. The targeting levers come from LINE’s own data: age and gender estimates, region, OS, and interest segments built from in-app behavior. You can also build custom audiences from your own user lists and retarget from there.
The format that has no Western equal is the Official Account. Brands run an Official Account as a follower base, then push messages, coupons, and rich menus to people who have added them. An install campaign can move a user from an ad to an Official Account follow, then convert later through a message. That 2-step path costs more setup work than a direct install ad, but it gives you a channel you own inside the app, which matters when the platform controls everything else.
LINE Stickers are the other lever. Sponsored sticker sets get distributed for free, users add them to keep using them, and the act of adding often requires following the brand’s Official Account. It is a discovery mechanism with no direct analog in Meta or Google.
Zalo in Vietnam
Zalo serves an estimated 75 million-plus users in Vietnam, and its ad system, Zalo Ads, is younger and thinner than LINE’s or WeChat’s. That cuts both ways. The targeting options are simpler: location down to province, age, gender, and platform, with fewer interest segments to work with. CPMs tend to run lower than in Japan or Korea, partly because the ad market is less crowded.
Zalo Official Accounts work like LINE’s, letting a brand message followers and run mini-programs called Zalo Mini Apps. The practical note for UA teams: creative copy has to read as local Vietnamese, not translated. Zalo’s user base skews toward people who do not use Meta as a primary surface, so a campaign that only runs on Facebook in Vietnam misses a large share of reachable users.
WeChat in China
WeChat is the strictest of the 3. China requires an ICP filing and a domestic business entity before you can run most ad inventory or publish a Mini Program, so the entry cost is regulatory before it is creative. WeChat Ads run through Moments (the social feed), Official Accounts, and Mini Program placements, and the review process for creative and landing pages is tighter than anything on the open web.
Mini Programs change the install question entirely. A WeChat Mini Program runs inside WeChat with no App Store download, so for some products the goal is not an install at all but a Mini Program open and retain. If your product can live as a Mini Program, you may skip the store. If it cannot, you are routing users from a closed ecosystem to an Android store outside WeChat, since there is no unified Google Play in China, and attribution across that gap is harder than in any other market here.
What to carry between them
The shared lesson is that the Official Account or equivalent owned channel is the asset, and the ad spend feeds it. The differences are where budget goes wrong:
- Targeting depth: LINE and WeChat expose richer segments; Zalo gives you geography and demographics and little else.
- Creative review: WeChat rejects creative and landing pages that LINE would clear without comment.
- Install vs. open: WeChat Mini Programs can replace the install; LINE and Zalo still send users toward a store in most cases.
Build the campaign for the platform’s rules first. Reuse what travels, which is mostly the owned-channel structure, and rebuild what does not, which is most of the targeting and all of the creative.
What this changes for your UA measurement and budgets
When a meaningful share of installs starts inside a closed app, your measurement stack inherits that app’s boundaries. The signals you rely on outside these markets, click IDs that survive the trip to the store, deep links that open a specific screen, postbacks that fire on a known schedule, all behave differently once a messaging platform sits between the ad and the install. Plan the measurement before you plan the spend, because the platform decides what you get to see.
Attribution inside a walled garden
Most MMPs (AppsFlyer, Adjust, Singular) read attribution from a click or impression that carries an identifier to the store. Inside KakaoTalk, LINE, WeChat, and Zalo, part of the journey happens in the platform’s own surfaces, so the identifier you depend on may not pass through cleanly. A shared link, a gift, or a mini-program tap can produce an install the MMP records as organic, because the referring signal stayed inside the platform.
2 practical effects follow. First, your organic count is inflated by installs that the messaging layer drove but did not tag. Second, the platforms run their own conversion reporting, and that report rarely lines up with your MMP. You end up reconciling 2 numbers that count different events. Treat the platform’s figure as the spend-side truth and your MMP as the retention and revenue truth, then reconcile on a known cohort rather than trusting either number alone.
Deep links that open a screen, not a store
Deep links carry more weight here than in store-first markets, because the install often skips the store entirely. A WeChat mini-program or a KakaoTalk channel link can drop a user into a specific flow without a store visit. That removes a measurement checkpoint you usually rely on, and it also removes a re-engagement path you may have built around store re-acquisition.
Build deferred deep links that survive the install and route the user to the screen the ad promised. If a user taps a gifting link for a specific in-game item, the link should open that item, not the home screen. The platforms support this, but each exposes its own link format and its own rules for what data the link can carry, so the implementation differs per market.
SKAN-style limits and the China exception
On iOS, SKAN constraints still apply in Korea, Japan, and Vietnam, so the conversion-value model you use elsewhere carries over. The difference is the placement: when the impression happens inside a messaging app rather than a self-attributing network you already integrate, you need to confirm the platform forwards a valid SKAN postback at all. Some surfaces do not.
China is a separate case. Apple’s SKAN reporting is not the relevant frame on Android, and most Chinese Android distribution runs outside Google Play through local stores and WeChat itself. Your measurement there depends on WeChat’s own analytics and the local store’s reporting, not on the MMP setup you use for the rest of Asia.
Where the budget should move
Once you accept that the messaging layer carries the channel, the spend mix follows:
- Hold back a test budget per platform before you scale, because cost per install on a Biz Board slot and on a LINE ad differ, and neither maps to your Meta CPI.
- Fund the integration work, deep links, postback validation, reconciliation, before you fund media. A campaign you cannot measure scales your blind spots.
- Set a separate target for messaging-sourced installs rather than rolling them into a blended number that hides which channel paid off.
The aim is a budget that reflects how installs reach you in each country, not a single allocation copied from a store-first market.
A practical entry plan for messaging-led markets
The mistake most teams make is treating “Asia” as one campaign. KakaoTalk, LINE, Zalo, and WeChat run separate ad systems, separate creative rules, and separate attribution behavior. A sequence that picks one market at a time keeps your test budget small and your read on the data clean.
1. Pick the platform by where your app already fits
Start with the country, not the channel. If your title is a casual game with social mechanics, Korea and KakaoTalk’s games tab reward sharing and gifting. If it depends on payments or local services, WeChat mini-programs in China and Zalo in Vietnam sit closer to that behavior. Map your retention loop to the surface that already hosts similar apps, then commit to one market for the first test. Spreading a small budget across 4 ecosystems gives you 4 underpowered tests and no clear signal.
2. Rebuild creative against local norms before spending
The creative that works on App Store Search Ads or Google UAC rarely transfers. Inside these apps your ad competes with chat content, so the rules differ:
- Biz Board and chat-list slots reward one image and one line of copy. Treat them like a push notification, not a 15-second spot.
- LINE tends to favor character-led and sticker-adjacent creative, because users already associate the app with that visual style.
- WeChat mini-programs need a working entry experience inside the shell, not just a tap that opens a store page.
Produce at least 3 creative variants per surface and localize the copy with a native speaker, not a translation tool. Tone and reading order carry weight in these markets, and a literal translation reads as foreign.
3. Read early signals without scaling
Closed ecosystems limit the attribution detail you get, so design the test to read on the signals you can actually see. Set a fixed budget cap per market for the first 2 weeks and watch:
- Cost per install per surface, tracked separately for each path. Biz Board and the games tab will not share a cost profile, and averaging them hides the channel that works.
- Day-1 and day-7 retention by source, since a cheap install from a gift link can retain very differently from an ad-driven one.
- Deep-link success rate, because broken deep links inside a messaging shell quietly drop users back to a generic screen and depress every downstream metric.
Hold scale until one surface clears your retention and cost thresholds in a single market. Then expand within that platform first, adding placements you have not tested, before opening a second country. This keeps each new variable isolated, so when a number moves you know what moved it.
The order matters: one market, localized creative, capped spend, then scale on proven retention. Skipping the sequence usually means paying tuition across 4 ecosystems at once.
