Treat every claimed advantage as a mechanism hypothesis把每项所谓优势都当作机制假说
For each claimed advantage, write a causal chain: a resource changes customer value or delivered cost; that affects choice, retention, price, or unit economics; rivals cannot quickly copy or neutralize it. Specify an observable measure, alternative explanation, matched cohort, and erosion signal. Pricing power, cost advantage, switching costs, network effects, licenses or intangibles, and scale economies require different evidence. Financial results may confirm economic consequences, but margins cannot identify the cause. Separate facts, calculations, assumptions, and judgments, and call the mechanism unproven when customer, competitor, or cohort evidence is missing.
对每项所谓优势写出因果链：某种公司资源或经营条件改变客户价值或交付成本；这种变化影响客户选择、留存、价格或单位经济；竞争者又无法迅速复制、绕开或抵消。随后指定一个可观察指标、一个可信的替代解释、一个匹配对照或 cohort，以及会削弱主张的侵蚀信号。价格权、成本优势、转换成本、网络效应、牌照或无形资产与规模经济需要不同证据。财务结果能够确认机制产生经济后果，但利润率本身不能识别原因。必须把报告事实、计算、假设与判断分开；若缺少客户、竞争者或 cohort 资料，就把机制保留为未证明。
Write the causal contract before reading the outcome在阅读结果前先写因果契约
| Causal link | Question | Evidence status |
|---|---|---|
| Resource or condition | What is distinctive: data density, process integration, license, know-how, location, or scale? | Usually disclosed fact plus operational detail |
| Customer or cost effect | What becomes better, cheaper, faster, safer, or harder to replace? | Measured outcome, not a management adjective |
| Economic capture | Does the company retain value through price, retention, share, or lower unit cost? | Calculation tied to a period and cohort |
| Resistance | Why can a rival not copy, subsidize, standardize, or route around it? | Competitive test with contrary evidence |
| Erosion condition | What observation would make the hypothesis weaker? | Precommitted falsification rule |
| 因果环节 | 问题 | 证据状态 |
|---|---|---|
| 资源或条件 | 独特之处是什么：数据密度、流程集成、牌照、诀窍、位置还是规模？ | 通常是披露事实加经营细节 |
| 客户或成本影响 | 什么变得更好、更便宜、更快、更安全或更难替代？ | 测量结果，而非管理层形容词 |
| 经济价值捕获 | 公司是否通过价格、留存、份额或更低单位成本保留价值？ | 绑定期间与 cohort 的计算 |
| 抵抗能力 | 竞争者为何不能复制、补贴、标准化或绕开？ | 包含反方证据的竞争检验 |
| 侵蚀条件 | 出现什么观察会削弱假说？ | 预先承诺的证伪规则 |
“Customers are sticky” skips the mechanism. A testable version is: “eleven embedded interfaces and two years of history make replacement disruptive, so renewal remains high after expiry despite a credible substitute.” It fails if customers dual-home, export cheaply, renew only because of penalties, or leave when standard connectors appear.
“客户黏性强”跳过了机制。可检验版本应是：“客户嵌入十一条接口与两年流程历史，替换连接会干扰协同，因此即使存在可信替代品，合同到期后仍保持高续约。”若客户普遍多平台并用、可低成本导出数据、只是因解约罚金续约，或标准接口出现后离开，该陈述就会失败。
This page tests whether a present mechanism is plausible. Calling it a durable moat requires longer evidence across competition, technology, regulation, and reinvestment. A successful one-period test is not a lifetime claim.
不要把当前证据升级成持久性本页只检验当前机制是否可信。要称为持久护城河，还需要穿越竞争、技术、监管与再投资的长期证据。单期检验通过并不是终身结论。
Give six advantage hypotheses six different failure tests让六类优势假说接受六种不同失败测试
| Hypothesis | Causal chain and observable evidence | Alternative explanation | Erosion test |
|---|---|---|---|
| Pricing power | Customer value supports realized price increases while comparable retention and usage hold | Inflation pass-through, better mix, or contracts delay churn | Renewal losses, discounting, usage contraction, or share loss after price |
| Cost advantage | A reproducible process lowers comparable service or production cost at equal quality | Understaffing, capitalized costs, lower service level, or favorable geography | Incidents, remediation, wage normalization, or rival cost convergence |
| Switching costs | Embedded workflows, data, retraining, and interfaces raise migration burden | Long contracts, penalties, inertia, or weak alternatives | Dual-homing, open exports, shorter migrations, and falling renewal |
| Network effects | Each relevant participant increases value for other participants in the same interaction network | Large markets attract both users and the vendor; correlation runs backward | Congestion, fragmented subnetworks, multi-homing, or no benefit from added nodes |
| Intangibles or licenses | Protected rights, approvals, trust, or know-how permit service others cannot match | A nonexclusive, expiring, or easily licensed permission adds little scarcity | Expiry, regulatory change, successful challenge, leakage, or substitute process |
| Scale economies | Fixed platform, security, or connector work spreads while marginal replication cost falls | Deferred maintenance or a simpler new market lowers cost temporarily | Customization resets cost, overhead grows, or small rivals match unit economics |
| 假说 | 因果链与可观察证据 | 替代解释 | 侵蚀测试 |
|---|---|---|---|
| 价格权 | 客户价值支持实际提价，同时可比留存与使用保持 | 通胀传导、组合改善或合同只是延迟流失 | 提价后续约损失、折扣、使用收缩或份额下降 |
| 成本优势 | 可重复流程在同等质量下降低可比服务或生产成本 | 人员不足、成本资本化、服务降级或地区成本有利 | 事故、补救、工资正常化或对手成本收敛 |
| 转换成本 | 嵌入流程、数据、培训和接口提高迁移负担 | 长期合同、罚金、惰性或替代品薄弱 | 多平台并用、开放导出、迁移缩短与续约下降 |
| 网络效应 | 同一互动网络中，每个相关参与者为其他参与者增加价值 | 大市场同时吸引用户与供应商，相关方向相反 | 拥堵、子网络割裂、多平台并用或新增节点无收益 |
| 无形资产或牌照 | 受保护权利、许可、信任或诀窍允许提供对手难以匹配的服务 | 非独占、即将到期或容易取得的许可缺少稀缺性 | 到期、监管变化、挑战成功、泄漏或替代流程 |
| 规模经济 | 平台、安全或接口固定工作被摊薄，边际复制成本下降 | 递延维护或新市场更简单造成暂时低成本 | 定制重新推高成本、总部膨胀或小型对手追平单位经济 |
Mechanisms may coexist, but one observation cannot be double-counted. High renewal after price can reflect value, switching, contract timing, or weak alternatives. Treat it as outcome evidence, then seek migration records for switching, node-driven usage for network effects, and matched unit costs for scale.
多种机制可以共存，但不能重复使用同一观察。提价后续约仍高，可能符合客户价值、转换成本、多年合同或缺少替代品等解释。先把它归为结果证据，再寻找机制专属证明：转换成本看迁移记录，网络效应看新增节点带来的使用改善，规模经济看匹配后的单位成本。一个指标不能承担六个结论。
Assemble TideGrid’s evidence without calling a winner整理 TideGrid 证据包，但先不宣布胜者
TideGrid Port Community Systems is fictional. Invented figures cover the 12 months ended December 31, 20X4 and 20X5. “Active” means at least five cross-organization workflows monthly. Density equals active ÷ eligible organizations within a port. Direct service cost includes implementation, support, and cloud delivery, excluding corporate overhead and stock compensation. ARR and retention are analytical measures, not accounting revenue.
虚构公司、定义与期间TideGrid Port Community Systems 是虚构港口社区协同软件公司。所有数据均为教学假设，期间为截至 20X4 与 20X5 年 12 月 31 日的十二个月。“活跃机构”指每月至少完成五次跨机构流程；参与方密度 = 港口活跃机构 ÷ 符合条件机构。直接服务成本包括实施与支持人工及云交付，但排除总部费用与股份薪酬。ARR 和留存是分析性订阅指标，不是会计收入。
| Observed item | 20X4 | 20X5 | Calculated movement |
|---|---|---|---|
| Revenue / gross margin | $100m / 68% | $125m / 72% | +25.0% revenue / +4.0 percentage points |
| Gross profit | $68m | $90m | +$22m / +32.4% |
| Mature-port participant density | 62% | 74% | +12 percentage points |
| Average net contract price, mature cohort | $50,000 | $54,000 | +8.0% |
| Gross revenue retention, mature cohort | 95.0% | 96.5% | +1.5 percentage points |
| Direct service cost per active organization | $16,000 | $12,500 | −21.9% |
| Connector reuse on latest port launch | 45% | 70% | +25 percentage points |
| Launch cash cost / duration | $12m / 18 months | $9m / 14 months | −25.0% cost / −22.2% duration |
| 观察项 | 20X4 | 20X5 | 计算变化 |
|---|---|---|---|
| 收入 / 毛利率 | 1.00 亿美元 /68% | 1.25 亿美元 /72% | 收入 +25.0% / 毛利率 +4.0 个百分点 |
| 毛利额 | 6,800 万美元 | 9,000 万美元 | +2,200 万 / +32.4% |
| 成熟港口参与方密度 | 62% | 74% | +12 个百分点 |
| 成熟 cohort 平均净合同价格 | 5 万美元 | 5.4 万美元 | +8.0% |
| 成熟 cohort 毛收入留存率 | 95.0% | 96.5% | +1.5 个百分点 |
| 每个活跃机构直接服务成本 | 1.6 万美元 | 1.25 万美元 | −21.9% |
| 最新港口上线的连接器复用率 | 45% | 70% | +25 个百分点 |
| 上线现金成本 / 时长 | 1,200 万美元 /18 个月 | 900 万美元 /14 个月 | 成本 −25.0% / 时长 −22.2% |
Gross profit is $100m × 68% = $68m and $125m × 72% = $90m. This is compatible with better price and delivery cost, but classification, capitalization, hosting mix, understaffing, or customer mix could move margin. The latest port may be cheaper because connectors were reused, or simply because local systems and labor were easier. Outcomes open mechanism tests; they do not close them.
毛利额分别为 1.00 亿 ×68%=6,800 万、1.25 亿 ×72%=9,000 万美元。结果符合更高价格与更低交付成本，却不能证明原因。收入分类、软件资本化、托管组合、人员不足与客户组合都可能改变毛利率。同样，最新港口上线更便宜，可能来自连接器复用，也可能只是当地海关系统与劳动力市场更简单。每个结果都开启机制检验，但不能自动结束检验。
Use cohorts and a port pair to challenge causality用 cohort 与港口配对挑战因果关系
| Design | Result | What it supports | What remains uncontrolled |
|---|---|---|---|
| Mature customer renewal cohort | Starting ARR $80m; churn $2.0m; contraction $0.8m; GRR = ($80m − $2.0m − $0.8m) ÷ $80m = 96.5% | Most starting revenue survived an 8% average price increase | Contract timing, discounts, usage, customer health, and absent alternatives |
| Interface migration sample | Median 11 live interfaces and 24 months of history; two completed exits took 7 and 10 months | Migration work exists | Two exits are selected observations; burden may reflect poor export tooling |
| New-port replication cohort | $9m and 14 months versus prior $12m and 18 months, with 70% connector reuse | A reusable implementation process is plausible | Port complexity, wage rates, service scope, and deferred remediation |
| License inventory | Seven port permissions; two renew within 24 months; none assumed exclusive | Permission is necessary in this example | Renewability and scarcity are not proven |
| 设计 | 结果 | 支持内容 | 仍未控制的因素 |
|---|---|---|---|
| 成熟客户续约 cohort | 期初 ARR 8,000 万；流失 200 万；收缩 80 万；GRR =（8,000−200−80）÷8,000=96.5% | 平均提价 8% 后，大部分期初收入仍保留 | 合同节奏、折扣、使用、客户健康与缺失替代品 |
| 接口迁移样本 | 中位数 11 条线上接口、24 个月历史；两次已完成退出耗时 7 与 10 个月 | 迁移工作确实存在 | 两个退出是选择性观察，负担也可能来自导出工具较差 |
| 新港复制 cohort | 最新为 900 万美元和 14 个月，之前为 1,200 万和 18 个月，连接器复用 70% | 可复用实施流程具有可信性 | 港口复杂度、工资、服务范围与递延补救 |
| 牌照清单 | 七个港口许可，其中两个在 24 个月内续期；假设均非独占 | 本案例中许可是必要条件 | 续期能力与稀缺性尚未证明 |
Anchor Port adopted TideGrid; similar Breakwater Port did not. Median coordination delay moved from 6.0 to 3.8 hours at Anchor and from 5.8 to 5.2 hours at Breakwater. The simple difference-in-differences is (3.8 − 6.0) − (5.2 − 5.8) = −1.6 hours. That association is consistent with an effect beyond the shared trend, but only if the ports would otherwise have followed parallel paths—an unverified assumption.
教学港口配对计算Anchor Port 采用 TideGrid，相似的 Breakwater Port 未采用。Anchor 的协调延迟中位数从 6.0 降至 3.8 小时，Breakwater 从 5.8 降至 5.2 小时。简单双重差分 =（3.8−6.0）−（5.2−5.8）=−1.6 小时。该关联符合超出共同趋势的效果，但前提是没有系统时两个港口本会沿平行路径变化；该假设尚未验证。
A stronger design checks several pre-periods, cargo mix, regulation, labor, weather, investment, and concurrent process changes. It asks whether rising density preceded more cross-party matches and shorter delays within the same port. Cross-sectional correlation is weaker because successful ports may attract both participants and TideGrid. Label evidence descriptive, associational, quasi-experimental, or randomized; a chart does not grant causality.
更强设计要检查多个前置期间、可比货种、监管、劳资行动、天气中断、投资与同期流程变化；还要检验机制：同一港口参与方密度上升后，跨机构匹配是否增加、延迟是否缩短？密度与表现的横截面相关更弱，因为成功港口可能同时吸引更多参与者与 TideGrid。必须标明证据是描述性、关联性、准实验还是随机实验，不能从一张图借用因果措辞。
Precommit to the signals that would reverse the conclusion预先承诺会推翻结论的信号
| Monitor | Current fictional observation | Adverse trigger | Mechanism challenged |
|---|---|---|---|
| Post-price renewal cohort | 96.5% GRR | Below 92% for two comparable renewals | Pricing power or switching costs |
| Participant density and workflow value | 74% density; delay association favorable | Density rises while matches, usage, or delay do not improve | Network effect |
| Direct service cost per active organization | $12,500 | Rises above $16,000 after normal staffing and incident remediation | Cost advantage |
| Port replication | 70% connector reuse; $9m launch cash | Reuse below 50% or next matched launch above $12m | Scale economy |
| Data portability | Median exit sample 8.5 months | Open standard reduces matched migration below 4 months | Switching cost |
| Permissions | Two renewals due within 24 months | Nonrenewal, open access, or easy rival licensing | License or intangible position |
| 监测项 | 当前虚构观察 | 不利触发器 | 被挑战机制 |
|---|---|---|---|
| 提价后续约 cohort | GRR 96.5% | 连续两个可比续约批次低于 92% | 价格权或转换成本 |
| 参与方密度与流程价值 | 密度 74%，延迟关联有利 | 密度上升而匹配、使用或延迟不改善 | 网络效应 |
| 每活跃机构直接服务成本 | 1.25 万美元 | 恢复正常人员与事故补救后高于 1.6 万美元 | 成本优势 |
| 港口复制 | 连接器复用 70%，上线现金 900 万美元 | 复用低于 50%，或下一可比上线超过 1,200 万美元 | 规模经济 |
| 数据可携带性 | 退出样本中位耗时 8.5 个月 | 开放标准让匹配迁移缩短至 4 个月以下 | 转换成本 |
| 许可 | 两个许可将在 24 个月内续期 | 不续期、开放准入或对手轻易获得许可 | 牌照或无形资产地位 |
These thresholds are analyst assumptions, not company guidance or universal rules; contrary evidence must reduce confidence. TideGrid’s fictional top five customers are 28% of revenue and the largest is 8%. Retention may reflect switching costs or bargaining by powerful buyers. Contract duration, termination rights, concessions, and win-loss data distinguish the explanations.
以上阈值是分析假设，不是公司指引或通用规则。其作用是建立纪律：反方证据出现时，假说置信度必须下降。还要监测客户集中度——TideGrid 虚构前五大客户占收入 28%，最大客户占 8%。高留存可能代表转换成本，也可能只是少数强势买方集中谈判。合同期限、终止权、价格让步与竞标输赢数据才能区分这些解释。
Finish with an evidence grade, not an advantage label以证据等级收尾，而不是优势标签
In the fictional packet, revenue, margin, contracts, concentration, licenses, and segment totals are treated as source facts.
报告事实在虚构证据包中，收入、毛利率、合同、集中度、许可与分部合计被视为来源事实。
Growth, gross profit, GRR, unit cost, launch changes, density, and the −1.6-hour port-pair estimate reproduce from stated inputs.
计算增长、毛利额、GRR、单位成本、上线变化、密度与 −1.6 小时港口配对估计均可按输入复算。
Metric definitions, port comparability, parallel trends, thresholds, cost scope, and counterfactuals are invented.
假设指标定义、港口可比性、平行趋势、阈值、成本范围与反事实均为虚构假设。
Mechanisms remain provisional without rival economics, complete churn reasons, matched launches, and post-standard migration evidence.
判断与缺口缺少对手经济性、完整流失原因、匹配上线与标准出现后迁移证据，机制只能保持暂定。
A proportionate conclusion reads: “TideGrid has evidence compatible with pricing, switching, network, cost, and replication mechanisms. The 8% price increase and 96.5% GRR support customer value; interface history supports migration burden. Density and the −1.6-hour port-pair estimate remain associational because parallel trends are unverified. Lower service and launch cost fit reuse, subject to quality and complexity controls. Nonexclusive permissions enable operation but do not establish scarcity. Revenue and margin confirm favorable outcomes, not cause. Confidence should fall if the stated renewal, portability, density-value, cost, replication, or license triggers fail.”
与证据相称的结论可以写成：“TideGrid 展现出符合价格、转换、网络、成本与复制机制的证据，但强度不同。成熟 cohort 提价 8% 后 GRR 为 96.5%，支持客户价值；接口历史支持迁移负担；密度与 −1.6 小时港口配对估计只属于关联证据，因为平行趋势未验证。更低服务和上线成本符合复用机制，但受服务质量和港口复杂度控制。非独占许可允许运营，却尚未建立稀缺性。收入增长与毛利率扩张确认经济结果有利，不能证明原因。若声明的续约、可携带性、密度价值、成本、复制或许可触发器失败，机制假说应被削弱。”
Upload evidence that can contradict the advantage story上传能够反驳优势叙事的证据
Upload annual and interim filings; business, competition, and risk sections; segment, product, geography, major-customer, and contract notes; KPI definitions and histories; implementation, support, cloud, capitalization, and incident costs; license terms; win-loss records; competitor disclosures; and management presentations. Include restatements and definition changes.
Ask Stock Explained to map each mechanism to source, page, note, table, period, cohort, numerator, denominator, currency, unit, and evidence status. Separate causal chain, observation, alternative, and erosion trigger. Open every original locator. Recalculate retention and unit costs, reconcile segments, confirm contract and license dates, inspect exclusions and survivorship, and compare risks with favorable presentations. Keep missing competitor or counterfactual evidence unresolved; never convert association into causation.
上传完整年度与中期披露、业务与竞争章节、风险因素、分部/产品/地区/大客户附注、客户集中度、收入确认政策、合同期限/续约/终止/折扣/服务等级明细、KPI 定义及历史表、实施/支持/云服务/资本化/事故成本、牌照条款和续期日、竞标输赢记录、竞争者披露与管理层演示，并纳入重述和定义变化。
要求 Stock Explained 把每项机制映射到来源、页码、附注、表格、期间、cohort、分子、分母、币种、单位与证据状态，并分别输出因果链、支持观察、替代解释与侵蚀触发器。随后在上传原件中逐个打开定位，重算留存和单位成本，把分部调节到合并金额，确认合同与牌照日期，检查排除项和 cohort 幸存者偏差，并将风险披露与有利演示对照。缺失的竞争者或反事实证据必须保持未解决，不能让生成摘要把关联升级为因果。
Upload contracts, KPI histories, segment notes, risks, and competitor evidence; require page-level support and verify every causal link in the originals.
上传合同、KPI 历史、分部附注、风险与竞争者证据，要求页级支持，并在原件中核验每个因果环节。
Open Stock Explained打开“一眼看懂这只股票”Questions after the mechanism tests are written写完机制测试后的常见问题
No. Margins are outcome evidence. They may reflect price, cost, mix, scarcity, accounting classification, underinvestment, or a cycle. Identify a mechanism, compare matched cohorts or units, and test whether the outcome survives rival response and normalized spending.
高利润率能证明竞争优势吗？不能。利润率是结果证据，可能来自价格、成本、组合、短缺、会计分类、投入不足或周期。应先识别机制，再比较匹配 cohort 或单位，并检验结果能否穿过竞争回应与正常化投入。
A network effect raises participant value when relevant participants join; scale lowers the provider’s unit cost as activity spreads fixed work. TideGrid needs user-value evidence for the first and matched service or launch costs for the second. Both may occur, but one does not prove the other.
如何区分网络效应与规模经济？网络效应是在相关参与者加入时提高参与者价值；规模经济是在固定工作被更多活动摊薄时降低供应商单位成本。TideGrid 前者需要用户价值证据，后者需要匹配服务或上线成本；两者可以并存，但不能互相证明。
No. Retention can reflect satisfaction, contracts, discounts, inertia, weak rivals, or switching burden. Review expiry cohorts, post-price renewals, termination rights, completed migrations, dual-homing, exports, and stated churn reasons before selecting the mechanism.
高留存能证明转换成本吗？不能。留存可能来自满意、合同、折扣、惰性、对手薄弱或迁移负担。选择机制前，应检查到期 cohort、提价后续约、终止权、已完成迁移、多平台并用、导出能力与流失原因。
Use a fixed entry rule, period, exposure age, denominator, and treatment; include exits rather than only survivors. Check customer mix and concurrent changes. A matched control and several pre-periods improve inference, but observational cohorts still require explicit causal limits.
什么让 cohort 比较更可信？使用固定进入规则、期间、暴露年龄、分母与处理方式，把退出者纳入而非只看幸存者；还要检查客户组合与同期变化。匹配对照和多个前置期间能改善推断，但观察性 cohort 仍需明确因果边界。
Reject or downgrade it when its precommitted mechanism-specific trigger fails, a better alternative explains the results, or replication cannot survive matched controls. Missing evidence is not positive evidence. A hypothesis can remain plausible without being established.
何时应拒绝竞争优势假说？当预先承诺的机制专属触发器失败、更好的替代解释出现，或复制结果无法通过匹配控制时，应拒绝或降级。缺失证据不是正面证据；假说可以仍然可信，却不能被视为已成立。
Primary sources for business, risk, segment, and contract evidence业务、风险、分部与合同证据的一手来源
The sources support where to locate business, competitive-condition, risk, segment, major-customer, and customer-contract evidence. They do not prescribe the fictional metrics or establish causal effects. Exact disclosures depend on framework, issuer status, materiality, jurisdiction, contract, filing type, and effective date. They do not validate TideGrid.
This page is educational, not economic, statistical, accounting, legal, regulatory, operational, or individualized investment advice. Mechanism evidence, retention, density, margin, and quasi-experimental estimates cannot alone establish durability, valuation, or future returns.
以上来源支持定位业务、竞争条件、风险、分部、大客户与客户合同证据，但不规定虚构指标，也不建立因果效果。准确披露取决于会计框架、发行人身份、重要性、司法辖区、合同、申报类型与生效日期；这些来源不验证 TideGrid。
本页仅供学习，不构成经济、统计、会计、法律、监管、经营或个性化投资建议。机制证据、留存、密度、利润率与准实验估计不能单独确定持久性、估值或未来回报。