Bain & Company has published two findings about online commerce that cannot both describe one industry: e-commerce is not displacing physical retail, and Chinese online sales were set to triple inside three years. Both calls were correct. China and the West were therefore never running one curve at different speeds, and the formats Western retailers import from Shanghai arrive calibrated to a market they do not operate in. The error sits with the reader who files the two findings under the same heading. It also sits with the consultancies that have never had a commercial reason to object.

The Western half of the pair has aged close to plan, though not on the metric most retellings reach for. Darrell Rigby argued in August 2014 that American online growth had already halved from its early-2000s pace and that e-commerce was not eating retail. His headline number, that e-commerce would climb from 11% to about 18% of Forrester’s top 30 retail categories by 2030, was never a call on total retail sales; he put the all-retail baseline at 6% in 2014. The Census Bureau’s second-quarter 2026 release puts the seasonally adjusted all-retail share at 17.1%, a different denominator than Rigby’s forecast and a different measure than the one he staked his 18% on. What held up is the thesis underneath the number: slow, compounding growth at a moment when the consensus expected a rout, not a rout that arrived on schedule.

The China half aged differently, and not by a matter of degree. Serge Hoffmann, Bruno Lannes and Jessica Dai forecast in 2012 that Chinese e-commerce would triple to RMB 1.5 trillion within three years. The number was the title of the brief, offered as the outer edge of what a consultancy would say out loud. China cleared it roughly on schedule, then spent thirteen years making it look timid. Online retail sales reached RMB 15.97 trillion in 2025, more than ten times the figure the brief was heading toward.

The clearest evidence that these are two markets rather than two dates is where China’s share stopped. Online sales of physical goods reached 24.9% of Chinese retail in 2020; the two available data points show it climbing to just 26.1% by 2025. The channel kept growing in yuan, but it stopped taking real ground from stores. The American share sits nine points below that on its own headline measure, climbing slowly enough that Rigby’s 2014 curve still tracks it. A ceiling is a property of a market, not a milestone on a shared timeline.

The mechanism is visible in Bain’s own 2012 chart, one row below the forecast everyone quoted. Under the online-share line, the brief tracks the retail market those shares were measured against, growing from RMB 9 trillion in 2007 to a projected RMB 23 trillion by 2013. Chinese online commerce took share of a market that was doubling underneath it, which meant physical retail could lose share and still open stores. American e-commerce has had no such cushion, and every point it gains comes out of sales an incumbent already books. A penetration curve hides its denominator, which is why the two look comparable and are not.

A share taken from a doubling market and a share taken from a flat one look identical on a chart and mean opposite things on a rent roll.

The borrowing this licenses runs in both directions, and neither direction checks the denominator. Western retailers import formats assembled for China’s conditions without importing the conditions: the clearest case is AI assistants such as Quark and WeChat displacing brand websites as the interface consumers shop through, a shift we traced in March. The traffic the other way is stranger. The phrase now used to explain the American ceiling, the “natural limits of online penetration,” comes from Bain’s China Shopper Report, where it described Chinese grocery categories such as diapers, infant formula and skincare flattening at high penetration. This publication used it that way in July. A ceiling observed in four Chinese packaged-goods categories in 2019 is not evidence about American apparel in 2026, whatever it happens to predict correctly.

The strongest objection is that America is not on a different curve but on the same one, later. On this reading China’s climb from single digits to a quarter of retail is what the United States is doing more slowly, and the gap closes without anyone deciding anything. For that to hold, the American share would have to keep compounding through the high teens without decelerating, and the Census release gives the argument something to stand on: online sales grew 12.2% year over year last quarter against 6.7% for retail overall. Bain’s own October brief shows the operators splitting rather than converging on one playbook: Shein, TikTok Shop and Temu are exporting the online-native model into price-sensitive, mobile-first segments of the US market, a pattern the brief treats as a direct threat to incumbents, while JD.com bought the physical store estate of MediaMarkt and Saturn in Europe, cleared by the Bundeskartellamt last September, instead of trying to sell German consumer electronics online from nothing. Neither move shows the aggregate American share converging on China’s: one captures demand the incumbent retail base was not serving, the other bets that a market with a dense store network is entered through the stores rather than around them.

That leaves the buyer of the deck holding a test rather than a template. Before importing a Chinese format, a retailer can ask what the format was solving for: a retail market expanding underneath it, a store network too thin to serve the demand, a consumer already transacting inside a messaging app. Where those conditions are absent, the format is not early; it is answering a question nobody asked. The two Bain findings were never in conflict, and reading them as one industry has always been a choice rather than an oversight. Of everything a retailer could get right this year, unlearning that costs nothing.