Indonesia has already settled the question the rest of retail is still arguing about, and the answer is that winning discovery does not deliver the basket. Asked which platform they open most often to shop online, 47.5% of Indonesian internet users named Shopee against 33.4% for TikTok Shop, in a survey of 8,700 people the Indonesian Internet Service Providers Association conducted face-to-face across all 38 provinces between 1 February and 15 March. The convention is to read that gap as a countdown, with the content platform ascending and the marketplace in managed decline. The GMV describes something duller and more permanent.

The asymmetry registers only when TikTok Shop’s shopping reach is set beside TikTok’s audience reach. TikTok’s advertising reach in Indonesia covered 88.9% of adults at the end of 2025, per DataReportal’s Digital 2026 Indonesia. Nine Indonesian adults in ten sit inside the content layer; one in three treats it as the place they shop. TikTok Shop has owned Tokopedia, the country’s largest home-grown marketplace, since early 2024, so the shortfall cannot be charged to a cold start or thin distribution. What it measures is the conversion ceiling of a feed.

The growth that reads as conquest is mostly internal transfer. TikTok Shop’s Southeast Asian GMV more than doubled in 2025, and within that, Indonesia’s US$13.1 billion made it the platform’s second-largest market in the world after the United States, per Momentum Works figures reported by ContentGrip. Regional growth on that scale can mean a platform is winning new spending, or it can mean the same spending is being rearranged among fewer players. Indonesia’s entire platform e-commerce market grew only 2.2% that year, on the same firm’s accounting. Almost all of TikTok Shop’s gain therefore landed inside a market that barely expanded, which means the balance came off another platform’s ledger.

The ledgers it came off were not Shopee’s. Bukalapak stopped selling physical goods in February 2025, by which point the category had shrunk to under 3% of its revenue. Tokopedia, which ByteDance controls, spent the same year undergoing what Momentum Works calls GMV rationalisation, a shrinking of transaction value even as the app kept its audience. The two are not the same measure: in the APJII survey itself, Tokopedia’s share of adults naming it their most-accessed platform rose from 9.57% in 2025 to 11.2% in 2026, the opposite direction from its GMV. A rival left the field, and TikTok moved money from its own marketplace tab into its own video tab and booked the transfer as share gain.

Shopee’s own disclosures show a basket that never moved. GMV grew 28.4% year on year in the second quarter of 2026 while gross orders grew 27.5%, the company told investors on 11 August. Order count and order value rising in near-lockstep is what it looks like when a marketplace adds transactions without buying them with discounts. Chief executive Forrest Li described the competitive situation on that call as “relatively stable at this point in time,” which is an odd way to characterise a market you are losing.

The more awkward fact for the substitution story is that Shopee has taken discovery back. Live streaming and short-form video now generate more than a quarter of Shopee’s physical-goods orders across Southeast Asia, Li said on the same call, with those orders up more than 50% year on year. Shopee’s defensible assets were always the stored card, the voucher balance, the saved address, the returns history and the rider network, and video turned out to be a feature that bolts onto all of them. Bolting a marketplace onto a feed is the harder build, and Indonesia has now priced both.

The feed won the decision, the marketplace kept the money, and nearly three years of trying has not moved either boundary.

The strongest case against this reading is a trend line. TikTok Shop’s share of Indonesians naming it their main shopping platform climbed about six points in a year, while Shopee’s fell by roughly the same margin against the 2025 edition of the same survey. If both slopes hold, the crossover arrives before the decade turns, and a settled structure is merely a photograph taken mid-overtake. For that reading to be right, one condition has to hold: TikTok Shop’s next tranche of GMV has to come out of Shopee’s basket rather than out of Tokopedia’s and Bukalapak’s. It has not yet, and the donors who funded the first phase are spent. Bukalapak has no physical goods left to lose, and Tokopedia cannot be rationalised twice.

For brands selling into Indonesia, the consequence is that the media budget and the commerce budget now have different addresses. Content investment belongs where the decision is made, in a feed that reaches almost every adult in the country; merchandising, pricing, assortment and service investment belongs where the transaction clears, which is still Shopee. Running the two as a single channel strategy, with one team and one set of targets, reliably produces a brand that is visible and unprofitable. The failure is organisational before it is commercial, because nobody in the building is measured on the handoff between the platform that persuades and the platform that charges.

Indonesia is the oldest version of this market, which makes it the most instructive. We argued in March that Southeast Asia’s video-native commerce model was the capability Western retail had failed to build, and the Indonesian data narrows that claim usefully. What exported was the discovery engine. Retailers in the US and Europe are not choosing whether the feed takes discovery, because it has taken it in every market where it has been allowed to try. They are choosing whether anyone else still holds the stored card, the address book and the returns history when it does, and Indonesia is the evidence that this is the half nobody can take from you.