From scattered PDFs to a Shopify checkout, in that order从零散的 PDF，到 Shopify 收银台，顺序不能反。

Why Stage 2 waited第二阶段，为什么要等？

The instinct, once the founder decided to fix this properly, was to move fast on everything at once: get the paperwork sorted and get a Shopify store and an Amazon listing built in parallel, since the brand assets weren't the hard part. We pushed back on the sequencing, not the ambition.创始人下定决心把这件事彻底解决之后，本能的想法是全面提速：一边整理文书，一边同步搭建 Shopify 独立站和亚马逊 listing，反正品牌资产不是难点。我们没有反对这份雄心，反对的是顺序。

Under the FSVP rule — the FDA's Foreign Supplier Verification Program, which makes a U.S. food importer responsible for confirming its overseas supplier meets U.S. safety standards — it's the U.S. importer, not the foreign factory, who is legally obligated to hold hazard-analysis and supplier-verification records for each product. In practice, that means no importer will accept the FSVP relationship without a factory-side evidence file solid enough to rely on, so a thin file is not a paperwork inconvenience.根据 FSVP 规则——FDA 的"境外供应商核查计划"，要求美国食品进口商核实自己的海外供应商是否符合美国食品安全标准——负有法律义务、必须为每一款产品保存危害分析和供应商核验记录的是美国进口商，不是境外工厂。这在实践中意味着：如果工厂一侧没有一份足够扎实、能让进口商放心依赖的证据文件，就不会有进口商愿意承担这层 FSVP 关系——所以文件单薄从来不是个文书上的小麻烦。

It's the reason a real buyer stays unreachable regardless of how good the direct-to-consumer (DTC) site looks. Building a storefront before that file existed would have meant marketing a product that no compliant North American buyer, wholesale or marketplace, could stock yet.这正是为什么哪怕独立站做得再漂亮，真正的买家也依然联系不上你。如果在这份资料齐备之前就先搭好门店，等于是在推广一款批发或线上市场里，任何合规的北美买家当下都还上不了架的产品。

So the rule we set going in was plain: Stage 1 first, evidenced, not assumed complete. Stage 2 opens only once that evidence is real. Everything downstream, the store, the listing, the ad spend, waited on that gate.所以我们一开始就定下了明确的规则：第一阶段先行，要有证据，而不是假定齐备。只有证据真正齐备，第二阶段才会开启。后面的一切——独立站、listing、广告投放——都要等这一关。

Building the file that could survive review搭建一份经得起审核的资料

The Readiness Diagnostic came first and it wasn't a green light. It was a scored gap list against the things a buyer's file actually gets checked against: FDA registration status, U.S. Agent, GFSI certification posture, whether the contaminant panel was mapped to the right thresholds, label completeness, and how much of the documentation existed only in Chinese.准备度诊断是第一步，它给出的不是"可以了"的信号，而是一份打分过的缺口清单，对照买家资料实际会被审核的那些项目：FDA 注册状态、美国代理人、GFSI 认证情况、污染物检测是否对照了正确的限值、标签是否完整，以及有多少文件只以中文形式存在。

From there, Easterlies coordinated the specialists the regulated steps required. An FDA U.S. Agent service. A GFSI-accredited certification body, with SQF (Safe Quality Food, a widely recognized food-safety certification scheme) as the target scheme. A US-accredited lab to re-run the heavy-metal and pesticide panel against FDA action levels (the maximum amount of a contaminant like lead the FDA allows in a food) and EPA pesticide tolerances (the maximum pesticide residue the EPA allows on a crop). Separately, the same panel was checked against California's Prop 65 lead threshold — a stricter, standalone limit that requires a warning label if crossed, regardless of whether the product is otherwise legal and safe under the federal rules above. That threshold sits low enough that tea and matcha commonly trigger a warning regardless of whether the product carries any elevated risk.在此基础上，Easterlies 对接了受监管环节所需要的专业机构：一家 FDA 美国代理人服务机构；一家 GFSI 认可的认证机构，目标体系定为 SQF（Safe Quality Food，一套被广泛认可的食品安全认证体系）；一家美国认证实验室，按 FDA 行动限值（FDA 允许食品中存在某种污染物如铅的最高量）和 EPA 农药残留容许量（EPA 允许农作物上某种农药残留的最高量），重新做一轮重金属和农药残留检测。此外，同一轮检测也对照了加州 Prop 65 铅阈值——这是一条更严格的独立限值，只要超过就必须附警示标签，无论产品在前面这些联邦规则下是否本就合法安全。这个阈值低到无论产品本身风险高低，茶叶和抹茶普遍都会触发警示。

Easterlies didn't become the FSVP importer, the U.S. Agent, the certifying body, or the lab. That coordination is a distinct job from doing the regulated work, and keeping that line clear was part of what made the file credible.Easterlies 没有变成 FSVP 进口商、美国代理人、认证机构或实验室。协调对接和亲自完成受监管的工作是两件不同的事，守住这条界限，正是这份资料能获得信任的原因之一。

Everything that came back fed into the Buyer-Readiness Dossier: hazard-analysis inputs written in FSVP-relevant language, an English spec sheet, an allergen statement and Nutrition Facts panel routed to a labeling specialist for sign-off, contaminant results mapped to the right thresholds, and the SQF audit trail once it existed. Alongside it, Buyer Standards Mapping separated what a natural-channel distributor actually requires from what a private-label CPG account requires, so cert spend stopped chasing certifications nobody downstream was asking for.所有回收的结果都汇入了买家资料包：用 FSVP 相关语言写成的危害分析输入项、英文规格表、送交标签专业人员签字确认的过敏原声明和营养成分表、对照正确限值的污染物检测结果，以及后续形成的 SQF 审核记录。与此同时，买家标准梳理把天然渠道经销商的真实要求，和自有品牌食品企业客户的真实要求区分开来，让认证方面的花费不再盲目追逐下游根本没人要求的认证。

Roughly five months in, FDA registration was active with a named U.S. Agent, the SQF certification had passed, and the contaminant panel came back clean against U.S. and California thresholds. The stalled distributor questionnaire went back out, complete this time, and came back marked under review by the buyer's own procurement team. That's a reviewer's judgment that the file is usable, not a placement, and we were careful not to let it read as more than that.大约五个月后，FDA 注册生效，在案美国代理人到位，SQF 认证通过，污染物检测结果对照美国和加州限值均为清洁。原本停滞的经销商问卷被重新完整提交，回复标记为"审核中"——这是审核方判断这份资料可用，不是拿到了订单，我们也很谨慎，没有让它读起来意味着更多。

The gate to Stage 2, and what it unlocked通往第二阶段的关卡，以及它解锁了什么

Substantially complete evidence was the bar, not a perfect file: registration and U.S. Agent active, certification passed or a dated audit scheduled against a clean pre-audit assessment, the contaminant panel clean against the right thresholds, and labels finalized with specialist sign-off. Once that bar was met, Stage 2 opened, and it opened as a validation question before it opened as a build.标准是"证据基本齐备"，不是一份完美无缺的资料：注册和美国代理人已生效，认证已通过或已排定审核日期且预审无重大缺口，污染物检测对照正确限值为清洁，标签已由专业人员签字定稿。达到这个标准之后，第二阶段才开启——而且它首先是以一个验证问题的形式开启，而不是直接开工建站。

A 30 to 45 day sprint tested positioning, price band, and channel fit for a single consumer-facing SKU — short for stock keeping unit, meaning one specific product listing — before any inventory or ad budget moved. The model that came out of it wasn't flattering on its own. A China-origin SKU carries a real combined duty and tariff burden on top of freight and fulfillment — including the Section 301 tariffs the U.S. has applied to many Chinese-origin goods since 2018, on top of standard import duty. A typical DTC food-and-beverage acquisition cost can outrun a first order's margin before repeat purchase has a chance to even out the math. That's a genuine constraint on this category right now, not a hypothetical one, and it's the kind of number a founder needs before committing real spend, not after.一场 30 到 45 天的冲刺，在任何库存或广告预算投入之前，先针对单个消费端 SKU（库存单位，即一个具体的产品条目）验证了定位、价格区间和渠道适配度。冲刺跑出来的模型本身并不讨喜：中国原产的 SKU 在物流和履约成本之外，还背负着实打实的综合关税负担——包括美国自 2018 年起对许多中国原产商品加征的301 条款关税，叠加在常规进口关税之上。典型的快消食品饮料获客成本，往往会在复购机会真正拉平账目之前，就吃掉首单的利润空间。这是这个品类当下真实存在的约束，不是假设出来的，也是创始人需要在真金白银投入之前而不是之后知道的数字。

The sprint's go signal was conditioned on clearing that math, specifically on the assumption that acquisition cost gets recovered across repeat orders rather than on the first sale alone. With that condition met, the Shopify build went ahead: storefront, content, and an email and SMS foundation. The Amazon build ran in parallel, Brand Registry setup, an FBA-compliant listing, and listing assets that carried the Prop 65 and Nutrition Facts language the compliance work upstream had already produced.冲刺给出的"继续"信号是有条件的，具体条件是：获客成本需要靠复购订单摊薄回本，而不是指望首单就打平。在这个条件被满足之后，Shopify 建站才启动：独立站、内容、以及邮件和短信基础设施。亚马逊这边同步推进——品牌注册、符合 FBA 要求的 listing，以及承载了上游合规工作早已产出的 Prop 65 和营养成分表信息的 listing 素材。

What closed, and what we won't claim到底落地了什么，我们又不会声称什么？

By month ten, the storefront was live with its email and SMS infrastructure running, and the Amazon listing was live under Brand Registry, FBA-compliant and carrying the label elements built earlier in the arc. The first 60 to 90 days of funnel and marketplace data came back inside realistic ranges for the category, reported as what it was rather than dressed up as a bigger result than the numbers supported.到第十个月，独立站已经上线，邮件和短信基础设施正常运行；亚马逊 listing 也已经在品牌注册下上线，符合 FBA 要求，带着这条链路更早阶段就做好的标签要素。最初 60 到 90 天的转化漏斗和市场数据，落在这个品类合理的区间之内，我们如实报告，不会把它包装成数字本身撑不起的更大结果。

We're not going to tell you this produced a specific revenue number, a guaranteed placement, or a ranking outcome, because none of that happened and claiming it would be dishonest. "Funded DTC" here describes the scenario, a factory that earned its way to an owned storefront and a marketplace presence, not a claim that money was raised. What we can stand behind is narrower and, we think, more useful: every gate in this arc closed on real evidence before the next stage opened, and every go decision along the way is one a disciplined operator could defend on the numbers in front of them at the time, not on optimism about what might come next.我们不会告诉你这带来了具体的收入数字、有保证的渠道落地，或者排名结果，因为这些都没有发生，声称发生了就是不诚实。这里说的"完成融资的独立站"，指的是一种情景设定——一家工厂凭本事走到了拥有自己的独立站和市场入口这一步，不是在声称它融到了钱。我们能负责任地说的，范围更窄，但我们认为也更有用：这条链路里的每一道关卡，都是在真实证据齐备之后才打开下一阶段，路上的每一次"继续"，都是当时摆在面前的数字能撑得住的判断，而不是对未来的乐观期待。

Most of the ten months went to the readiness side, not because that's where the interesting work is, but because that's where the actual risk sat. The storefront and the listing were the easy part once the file underneath them could survive review.这十个月里，大部分时间都花在了准备度这一侧，不是因为那里最有意思，而是因为风险本就集中在那里。一旦底层的资料能经得起审核，独立站和 listing 反而是水到渠成的那一部分。