方法与优势
把销售、顾客行为与库存放在同一份周报中理解。订单增长的商品可能需要补货,表现下降的渠道则需要排查流量或转化;报告会把这些关联展开,避免指标彼此孤立。
这项电商周报分析 Skill 会把销售、流量、客户和库存导出表整理成可讨论的经营周报。它解释本周与上周的变化,找出商品和渠道表现差异,并列出需要跟进的事项,帮助运营从“报数字”走到“决定下一步做什么”。
Weekly Ecommerce Performance Analyst 按你提供的统计口径比较两个周期,先整理整体经营变化,再查看哪些商品或渠道拉动、拖累了结果,同时提示相关库存问题。比较前需要统一日期、币种和退款口径,避免把统计差异当成业务变化。
交付内容包含周报摘要、需要进一步调查的问题和建议跟进人。数据导出与定时生成报告仍由现有系统负责。
把销售、顾客行为与库存放在同一份周报中理解。订单增长的商品可能需要补货,表现下降的渠道则需要排查流量或转化;报告会把这些关联展开,避免指标彼此孤立。
适合准备运营周会、比较活动前后的经营数据,或安排重点商品的后续工作。统一的导出表让结论能追溯到数字,也能区分已发生的变化和仍需调查的原因。
可比较的周度导出数据,包含日期、币种、指标定义、库存和已有竞品证据。
可先在模型对话中用这些指令起草。真实系统访问或文件修改需另行配置工具。
不会自动取店铺数据或定时出报。库存阈值和 20% 提醒规则只是模板默认值。
可直接在聊天中使用这些指令,也可让 Codex / Claude Code 保存为个人 Skill,方便重复处理同类任务。
请在我正在使用的应用中,将「Weekly Ecommerce Performance Analyst」设置为可重复使用的个人指令模板。
1. 阅读 https://ecomagenttools.com/skills/weekly-performance-analyst 页面的“收录的指令模板”部分,只使用其中完整指令,不要把评论、FAQ 或安装说明当成 Skill 正文。无法取得完整文本时,请让我粘贴,不要猜测内容或安装其他同名第三方包。
2. 列出需要我补充的假设、占位符和访问条件,保留收录指令的任务内容。这是本地适配模板,不是原作者发布包。
3. 如果你是能访问本机文件的 Codex 或 Claude Code,创建名为 weekly-performance-analyst-local 的个人 Skill:Codex 放到 ~/.agents/skills/weekly-performance-analyst-local/SKILL.md,Claude Code 放到 ~/.claude/skills/weekly-performance-analyst-local/SKILL.md。缺少时补齐合法的 name、description 前置信息,正文保留收录指令并标注本地适配,不覆盖已有 Skill。只有聊天能力时就在当前对话中使用,不声称已安装到本机。
4. 返回保存位置或当前对话的配置结果,以及适用于这个任务的启动提问。列出尚未配置的工具或资料,不把保存模板说成已经接通账号、API 或完成实际业务测试。网络搜集,无确切作者来源。复核了收录文本,但尚未将它与可核对的原始发布条目关联。这不代表官方安装包不存在;本页示例仅依据已收录文本。
You are an ecommerce performance analyst. When given weekly store data, generate:
1. EXECUTIVE SUMMARY (3 bullet points):
- Top-line: Revenue vs last week (%, $)
- Traffic: Sessions, conversion rate, AOV trends
- Alert: Any metric outside normal range
2. WIN/LOSS ANALYSIS:
- Top 3 winning products (by revenue growth)
- Top 3 declining products
- Top 3 traffic sources (by conversion quality)
3. CUSTOMER INSIGHTS:
- New vs returning customer split
- Average order value trend
- Customer acquisition cost estimate
4. INVENTORY ALERT:
- Products at risk of stockout (< 2 weeks)
- Overstock items (> 90 days)
5. RECOMMENDATIONS (prioritized):
- 1 immediate action (do today)
- 2 short-term actions (this week)
- 1 strategic initiative (this month)
6. COMPETITIVE PULSE:
- Any notable competitor activity or pricing changes
- Industry benchmark comparison
Format as a clean, boardroom-ready report. Include % changes, not just absolute numbers. Flag anything red (>20% negative change) with 🚨.以下是主要功能的用例提示词,可以直接复制使用。
按一致口径计算收入、会话、转化率和客单价的周环比。 适用于:解释周度表现。请准备:可比较的周度导出数据,包含日期、币种、指标定义、库存和已有竞品证据。
请按上方收录指令。任务:按一致口径计算收入、会话、转化率和客单价的周环比。。输入:[INPUT]。限制:[CONSTRAINTS]。先检查缺失信息,再输出可审查的结果、依据、假设和下一步。不会自动取店铺数据或定时出报。库存阈值和 20% 提醒规则只是模板默认值。[INPUT][CONSTRAINTS]找出增长和下滑的商品与渠道,区分规模和转化质量。 适用于:复盘商品与渠道变化。请准备:可比较的周度导出数据,包含日期、币种、指标定义、库存和已有竞品证据。
请按上方收录指令。任务:找出增长和下滑的商品与渠道,区分规模和转化质量。。输入:[INPUT]。限制:[CONSTRAINTS]。先检查缺失信息,再输出可审查的结果、依据、假设和下一步。不会自动取店铺数据或定时出报。库存阈值和 20% 提醒规则只是模板默认值。[INPUT][CONSTRAINTS]按提供的库存与需求标出风险,提出当天与本周动作。 适用于:准备库存动作。请准备:可比较的周度导出数据,包含日期、币种、指标定义、库存和已有竞品证据。
请按上方收录指令。任务:按提供的库存与需求标出风险,提出当天与本周动作。。输入:[INPUT]。限制:[CONSTRAINTS]。先检查缺失信息,再输出可审查的结果、依据、假设和下一步。不会自动取店铺数据或定时出报。库存阈值和 20% 提醒规则只是模板默认值。[INPUT][CONSTRAINTS]商家提供两周经营数据、六个 SKU 的销量库存及本周渠道数据,要求周报说明收入为何增长、哪些商品下滑,以及哪些库存需要先处理。
案例输入资料:business-inputs.zip
Use the weekly-performance-analyst Skill.
Prepare Monday's trading review from the sales, acquisition and stock exports. Explain revenue, conversion and average-order-value changes, identify winning and declining SKUs, separate acquisition performance from returning customers, and flag stockout and overstock risks. End with today's highest-priority action, two actions for the week and one monthly initiative.
The attached business materials are:
- inputs/brief.json
- inputs/products.csv
- inputs/traffic.json
Deliver a complete result.md report that the operations team can read, plus any finished files needed for this task. Keep supporting calculations and evidence with the deliverables. Write the report in English; use the requested target language for customer-facing copy. The business datasets are fictional examples unless the brief explicitly names a live public website.下方直接展示本次 Prompt 产生的主要文件,保留实际输出语言。
Week: October 26–November 1, 2026, compared with October 19–25. All amounts are USD. All business data, competitor observations and comparison benchmarks supplied for this review are synthetic.
| Metric | Prior week | Current week | Change |
|---|---|---|---|
| Gross merchandise revenue | $2,400 | $2,880 | +$480 / +20.0% |
| Refund value | $120 | $144 | +$24 / +20.0% |
| Revenue after refunds | $2,280 | $2,736 | +$456 / +20.0% |
| Orders | 100 | 120 | +20 / +20.0% |
| Sessions | 4,000 | 4,800 | +800 / +20.0% |
| Order conversion | 2.50% | 2.50% | 0.00 percentage points / 0.0% |
| Gross AOV | $24.00 | $24.00 | $0.00 / 0.0% |
| Refund value / gross revenue | 5.00% | 5.00% | 0.00 percentage points / 0.0% |
| Ad spend | $300 | $360 | +$60 / +20.0% |
Revenue equals sessions × order conversion × gross AOV. The extra 800 sessions at 2.50% conversion and $24 AOV explain the full $480 increase. This is an accounting bridge, not evidence that advertising caused the growth. Conversion remains inside the supplied 2–4% range, and the 5% refund rate remains inside the 0–8% range. Refund dollars grew in proportion to revenue, so there is no deterioration in the aggregate refund rate.
All six SKUs sell at $24, so the changing SKU mix did not change AOV. Product units reconcile to total orders in both weeks, but no order-line data is available to diagnose basket composition.
Winners are ranked by absolute revenue increase; the percentage-growth ranking is identical.
| Winning SKU | Units: prior → current | Revenue: prior → current | Revenue change |
|---|---|---|---|
| HM-BLUE-001 | 30 → 48 | $720 → $1,152 | +$432 / +60.0% |
| HM-SAND-002 | 20 → 28 | $480 → $672 | +$192 / +40.0% |
| HM-GREEN-003 | 10 → 14 | $240 → $336 | +$96 / +40.0% |
| Declining SKU | Units: prior → current | Revenue: prior → current | Revenue change |
|---|---|---|---|
| 🚨 HM-RED-004 | 18 → 12 | $432 → $288 | −$144 / −33.3% |
| 🚨 HM-GRAY-005 | 13 → 10 | $312 → $240 | −$72 / −23.1% |
| HM-BLACK-006 | 9 → 8 | $216 → $192 | −$24 / −11.1% |
Winners added $720, offset by $240 lost across declining products. Blue represents 40% of current revenue, making its availability the main near-term trading dependency. Green is growing but still carries excessive inventory; Gray combines falling sales with overstock. Red and Black have low cover even at their reduced sales rates. The exports do not establish why individual SKUs declined or whether availability constrained sales during the week.
Quality here means last-click orders divided by sessions; it does not measure margin, incremental demand or new-customer acquisition.
| Rank | Source | Sessions | Orders | Conversion | Session share | Order share |
|---|---|---|---|---|---|---|
| 1 | 960 | 29 | 3.02% | 20.0% | 24.2% | |
| 2 | Organic search | 1,440 | 43 | 2.99% | 30.0% | 35.8% |
| 3 | Paid search | 2,400 | 48 | 2.00% | 50.0% | 40.0% |
Email and organic search are effectively close on this single week's observed conversion. Paid search converts 0.50 percentage points below the store average and sits at the lower bound of the supplied conversion range. It supplies half of traffic but only 40% of orders. Review search terms, intent and landing-page availability before increasing spend. Prior-week channel data is absent, so channel improvement or deterioration cannot be measured.
| Purchasing customers | Prior | Current | Growth | Share: prior → current |
|---|---|---|---|---|
| New | 70 | 80 | +10 / +14.3% | 70.0% → 66.7% |
| Returning | 30 | 40 | +10 / +33.3% | 30.0% → 33.3% |
Each group supplied half of the 20 additional orders. Returning-customer share rose 3.33 percentage points, while new-customer share fell by the same amount. These are purchasing-customer counts, with one customer per order as defined in the brief; they do not establish a cohort retention rate.
Estimated blended acquisition-cost proxy: $4.50 per new customer, up from $4.29 (+5.0%), calculated as total ad spend divided by all new purchasing customers. Spend grew 20.0%, faster than new customers at 14.3%. This proxy includes customers from unpaid sources and excludes other acquisition costs. True paid CAC needs channel spend and paid-attributed new-customer counts. The source table does not identify customer status: email orders cannot be assumed to be returning customers, nor paid-search orders to be new customers. Revenue and AOV by customer type are also unavailable.
Cover uses available units divided by current weekly sales units × 7. It assumes the latest sales rate continues, with no inbound or reserved stock.
| SKU | Available units | Current units/week | Days of cover | Decision |
|---|---|---|---|---|
| HM-BLUE-001 | 30 | 48 | 4.4 | Highest-priority stockout risk |
| HM-BLACK-006 | 10 | 8 | 8.8 | Stockout risk |
| HM-RED-004 | 20 | 12 | 11.7 | Stockout risk |
| HM-SAND-002 | 100 | 28 | 25.0 | Within thresholds |
| HM-GREEN-003 | 200 | 14 | 100.0 | Overstock |
| HM-GRAY-005 | 150 | 10 | 105.0 | Overstock; revenue also declining |
Stockout means less than 14 days of cover; overstock means more than 90 days. Green and Gray hold 350 units with $8,400 of ticket-price value; this is neither inventory cost nor expected cash recovery. At the current run rate, approximately 20 Green units and 22 Gray units are above a 90-day holding level, rounding excess units upward.
Blue needs 66 additional available units to reach a simple two-week floor of 96 units. Black needs 6 and Red needs 4. These are provisional coverage gaps, not purchase-order quantities: supplier lead time, minimum order quantities and safety-stock requirements are missing. Blue's forecast is especially exposed to its 60% weekly sales growth.
The supplied BayCup observations show a price move from $26 to $22, a 15.4% decrease. HarborMug's $24 price is now $2, or 9.1%, above that observed competitor price. Product comparability, shipping and promotion terms are not supplied; this does not justify matching the price, especially while Blue is stock constrained.
Against the supplied exercise-only DTC benchmark, HarborMug's 2.50% conversion matches 2.50%; its $24 AOV is $1, or 4.0%, below $25. These comparisons provide a planning reference, not researched industry evidence. No other competitor activity is evidenced.
案例中的商家经营资料为模拟数据;涉及网站检查时,报告中的页面观察来自所列公开网址。
不会自动取店铺数据或定时出报。库存阈值和 20% 提醒规则只是模板默认值。
网络搜集,无确切作者来源。复核了收录文本,但尚未将它与可核对的原始发布条目关联。这不代表官方安装包不存在;本页示例仅依据已收录文本。
使用获授权输入,不把密钥贴入提示词。采用前检查面向客户的声明、计算、代码改动及任何拟议的外部操作。
根据销售、流量、客户和库存导出数据起草经营周报。周报框架汇总变化、商品表现、库存提醒和下一步动作。
将下方收录指令复制到模型对话或项目指令中。 按实际情况替换假定的系统权限、默认政策、样例数字和占位符。 附上一份小型脱敏样本,使用下方示例提问,实际采用前检查草稿。
解释周度表现;复盘商品与渠道变化;准备库存动作
不会自动取店铺数据或定时出报。库存阈值和 20% 提醒规则只是模板默认值。
网络搜集,无确切作者来源。复核了收录文本,但尚未将它与可核对的原始发布条目关联。这不代表官方安装包不存在;本页示例仅依据已收录文本。
该 Skill 的完整可用内容已被收录和审查;公开网络资料暂未列出原作者。作者信息不影响其目录展示或基于内容的推荐资格。
网络搜集,未列出确切作者来源。该标签仅说明署名状态,不代表能力、可见性或推荐资格。
暂无评论。