THESIS: For a decade, creator content had one job: reach. The purchase happened somewhere else — a store, a site, a retargeting funnel — and content got credit for "awareness" while performance channels took the revenue. That era just ended, measurably. When TikTok Shop asked 17,000 users, 57.8% said they bought directly through the platform after discovering the product there — the full journey, content → discovery → trust → purchase, completed inside one feed session. In Brazil, that model grew 102× in average daily GMV in its first year. The strategic consequence for e-commerce brands: creator content is no longer the top of your funnel. It is the funnel — and it deserves to be built and measured like a sales channel, not a media experiment. But one creator posting one video is a lottery ticket, not a channel. A channel needs the four things this guide covers: the right creators tested systematically, content built for the purchase (not the like), posting volume that finds winners on schedule, and tracking that reads all the way to the sale.
CONTEXT: The verified Brazil numbers (TikTok's own first-year report, May 2025→May 2026): average daily GMV ×102; active affiliated creators ×46; daily livestreams ×20 and live-commerce GMV ×161; a LatAm record of $100K sold in a single broadcast; and the 57.8% direct-purchase figure from the 17,000-user in-app survey. Read those multipliers together and the mechanism is obvious: the GMV didn't grow 102× because ads got better — it grew because the creator base grew 46× and the content formats matured. Social commerce scales on creator supply and content volume, not on media budget. That's the machine to build — and everything below works whether the checkout is TikTok Shop itself, or your own store fed by shoppable content in any market where Shop hasn't launched yet.
Four parts, matching the guide's four promises.
═══
The principle: in social commerce the creator isn't your media placement — she's your salesperson. The selection logic follows: you're not casting for reach, you're casting for the ability to move product to a matched audience. Test wide, rank on sales, keep the sellers.
The test portfolio (don't pick one type — run all four in wave one):
The test discipline: 10–20 creators in wave one, each with her own affiliate link/code, 2–4 weeks, ranked on sales per 1,000 views — never on views, never on followers. Out of 10, expect 2–3 to sell above floor; you cannot predict which, so the test is the strategy. Keep the sellers, rotate the rest, and source wave two to look like wave one's winners.
═══
The principle: content that sells inside the feed has a different job than content that entertains. The viewer has to travel from "what's that" to "I trust this" to "I'm buying" in under a minute — so every video is a compressed sales conversation wearing native clothes. The formats that do it: