Cas intégré : vendre directement aux consommateurs à l'étranger

A Chinese supplier offers a three-person team exclusive access to one insulated tumbler. The team has no factory, no store and a budget that covers about six months of fixed costs. It can list on a marketplace tomorrow or build its own store in six weeks, and it has decided to sell at US$29 with free shipping in one overseas market. The question is not whether direct selling is fashionable. The question is which choices must be made before money is spent, and which numbers would show that the launch is working.

Start with the form (§2.2.4). The textbook lists five B2C operating forms: direct sellers, who buy and resell goods; online intermediaries, who do not own the product and match buyers and sellers for a commission; advertising-based B2C, which gathers traffic with free content and sells ads; community-based B2C, built on shared interests; and live-broadcasting B2C, which started in China and is spreading. The team is a direct seller because it takes title to inventory and carries the selling risk. Content and livestreaming are channels it may use, not a different form. The distinction matters because each form earns differently. An intermediary earns commission and needs liquidity on both sides; an ad-based site needs audience scale; a community needs repeated interaction; a live seller needs content and supply. A direct seller earns the difference between price and the full cost of the order. Mixing the labels produces bad plans, such as a three-person team deciding to become a platform because it has a website. It is still a direct seller with a storefront.

Before naming the brand, run the §2.2.6 checklist: research the product, market, audience, competition and traffic cost; choose a legal structure and obtain tax numbers, licences and permits; build a site with a payment gateway if you sell on your own domain; and market through as many channels as you can, starting small and narrow. The checklist is a risk register, not a queue. The team can prepare a listing while waiting for a licence, but it cannot collect money without a payment path, and it should not buy traffic before it knows the unit cost of a visit.

Mobile-first is the next constraint (§2.2.5). The textbook describes Wish and SheIn as cross-border apps operating only through mobile devices and summarises their advantages as precise customers, convenience and logistics. For this team, mobile-first is not building an app. It means three testable requirements: the price and buy button are visible above the fold; checkout takes no more than three steps; and the creative is vertical and loads quickly on a mid-range phone. A desktop page shrunk to a phone fails all three, and the failure appears as a low cart-to-order rate rather than as a design complaint.

Now price the order. The tumbler costs US$7.40 including packaging; inbound freight is US$1.60; last-mile delivery is US$3.80; platform commission is 8%, or US$2.32; payment fees are 2.5%, or US$0.73; and acquisition costs US$6.90 per order at the current conversion. Contribution before returns is 29.00 minus 22.75, or US$6.25. At a 6% return rate and US$9.00 cost per return, returns take US$0.54, leaving US$5.71 per order, or 19.7% of price. This is a teaching simulation, not a quote, but the structure is the point: marketing is the second-largest line after goods.

The funnel explains where that US$6.90 goes (§3.4.1). The textbook funnel runs from awareness through interest, consideration, intent and evaluation to purchase, and it is used to forecast sales and find stuck deals. For measurement, the team uses five observable stages: 1,200,000 impressions, 28,800 visits at 2.4%, 2,736 add-to-carts at 9.5%, 684 purchases at 25%, and 41 returns at 6%. Each rate answers a different question. A weak visit rate is a creative or targeting problem; a weak cart rate is a page, price or trust problem; a weak checkout rate is a shipping, payment or friction problem.

Diagnosis precedes spending. If visits doubled but the cart rate stayed at 9.5%, orders would double and contribution would rise, yet the underlying page problem would remain. If the cart rate rose from 9.5% to 12% at the same traffic, orders would reach 864 and net contribution would reach about US$4,934 a month. The team should fix the stage with the largest gap to a realistic benchmark, not the stage that is easiest to advertise. This is also why a single conversion rate is useless without its denominator.

Fulfilment is a second trade-off. Sending each order by parcel costs US$5.40 in inbound and last-mile freight and takes about ten to fourteen days. An overseas warehouse cuts the variable cost to US$4.50 and delivery to about six days, saving US$0.90 per order, but it requires 1,000 units of stock, or US$7,400 of cash. If a quarter of that stock is never sold and is written off at US$1,850, the saving needs about 2,056 orders to recover. The warehouse is cheaper per order only if demand is stable enough to sell the stock.

Returns and repurchase decide whether the contribution survives. A 6% return rate is tolerable only if the reasons are known: leakage, colour mismatch and transit damage require different fixes. Prevention through a leak-test video, daylight colour photos and drop-tested packaging is cheaper than a cross-border return. Repurchase for a durable tumbler comes from new colours, gift bundles and replacement lids, not from the same unit wearing out. The team should measure 60-day repeat purchase and treat it as a separate target from acquisition.

Finally, the channel choice. A marketplace charges 8% commission but brings search traffic; an own site avoids commission but must buy more traffic and carry higher fixed costs. In the simulation, the own site earns US$6.53 net per order against US$5.71, but adds US$1,800 a month in fixed cost, so it needs about 2,195 orders a month to break even. If owned channels cut acquisition from US$8.40 to US$7.20, the hurdle falls to about 890 orders. The recommendation must therefore be conditional: stay on the marketplace while the team has no owned audience and returns stay at or below 7%; test a warehouse only for the best-selling colour once weekly demand is stable; and revisit the own site when email and messaging lower acquisition enough for about 900 orders a month to break even. These thresholds are teaching numbers, not forecasts. The discipline is to write the condition, the metric and the review date before the launch, so that the next decision is based on evidence rather than enthusiasm.