DXP Video Tutorials & Playbooks
English learning series
Watch the new isA DXP learning series with English on-screen captions, chapter navigation, transcripts, and practical review points.
The DXP Video Hub compiles high-definition operational walkthroughs recorded in SweetNight production environments. Covering asset discovery, Active Copilot post-campaign review, What-If sandbox simulations, sales anomaly detection with human escalation, and advertising audit memos, each video provides hands-on instructions for operational leaders.
All tutorials feature precision timestamp chapters and compliance verification checklists.
5 Core Operational Video Courses
| Course Code & Title | Learning Objective & Operations Path | Key Deliverable | Duration & Chapters |
|---|---|---|---|
| 01 · Search Price Elasticity & Export Samples | Search elasticity models in the catalog, verify SLA contracts, and export compliant samples via Table Explorer. | Validated CSV export with batch tags and schema definitions | 08:45 (4 chapters) |
| 02 · Multi-Touch Attribution with Copilot | Provide multi-channel logs to Copilot, audit Shapley attribution calculations, and correct algorithmic drift. | Attribution log with explainable reasoning trail | 11:20 (5 chapters) |
| 03 · Sandbox What-If Simulation Boundaries | Adjust 5 macro sensitivity sliders in the Digital Twin to evaluate margin stress vs. live command dispatch boundaries. | Stress-test tolerance thresholds & optimal hedge parameters | 14:10 (6 chapters) |
| 04 · Detect Sales Anomalies & Escalate to Human | Apply anomaly detection algorithms to identify volume drop-offs and assemble standardized human review tickets. | Anomaly slice ticket with automated Feishu alert | 09:30 (4 chapters) |
| 05 · Operations Review & Audit Memo Preparation | Review multi-dimensional ad metrics, align time-zone cutoffs, and prepare audit-ready compliance memos. | Signed AEP audit memo package with BigQuery SQL | 12:05 (5 chapters) |
Interactive Video Player Hub
找到价格弹性数据产品，导出可核对的样本
Core tutorial excerpt
Goal: 开始分析前，需要找到价格弹性产品，理解记录结构，并取得可继续使用的数据样本
Result: 通过真实生产界面进入对应产品资产中心，检查字段，实际导出CSV和JSON；本次8行21列样本与服务端响应逐值一致。
找到价格弹性数据产品，导出可核对的样本
This lesson uses demonstration or sample data.
- 已完成查找、字段核对和当前8行样本交接；未完成AI分析、调价、决策执行、自动化或挖掘。
用 Copilot 复盘广告样本，核对计算并纠正结论
Core tutorial excerpt
Goal: 需要统一口径比较活动效率，并识别模型结论是否超出数据所能支持的范围
Result: 三条合成输入的比率与合计通过Decimal复算；实际追问纠正了净利润推断，最终结果交给人工核对。
用 Copilot 复盘广告样本，核对计算并纠正结论
This lesson uses demonstration or sample data.
- 独立对话中手动提供三条合成广告样本；未原生绑定前片价格弹性8行样本。ROAS/ACOS不能证明净利润或增量回报；追问给定数值后纠正了不受支持的推断，未执行预算、审批或自动决策。
从销量样本发现异常并交给人审
Core tutorial excerpt
Goal: 先发现异常，再核对计算，最后把待查问题交给人
Result: 标出异常样本，指出模型计算错误，并整理待人审问题
从销量样本发现异常并交给人审
This lesson uses demonstration or sample data.
- 仅分析公开合成销量样本；未下发业务动作或调用生产数据。
运营复核广告分析并形成补证备忘
Core tutorial excerpt
Goal: 同时核对广告指标和结论，整理成可核对的备忘
Result: 完成广告样本复盘，形成包含指标、公式与结论的备忘
运营复核广告分析并形成补证备忘
This lesson uses demonstration or sample data.
- 基于已有真实交互与合成数据重编；未新增模型请求或接入企业广告数据。
比较参数假设，理解推演与执行的边界
Supporting review and troubleshooting
Goal: 需要理解多项条件变化后方案可能面临的风险，并确认比较过程可以恢复起点
Result: 两组预设改变了页面输入与模拟结果；重置后六项输入和主要输出恢复到起始状态。
比较参数假设，理解推演与执行的边界
This lesson uses demonstration or sample data.
- 只比较页面预设的多项参数假设与模拟输出，并验证重置恢复；不作单变量因果归因，不代表模型训练、可信线上公式、策略执行或实际利润结果。
Course Chapters & Operational Playbooks
Course 01: Search Price Elasticity & Export Samples
- [00:00 - 02:15] Business Trigger: Raw material inflation signals prompt search for
SN-MAT-COOLNEST-12demand elasticity data. - [02:16 - 04:30] Catalog Query: Access
/dxp/marketplaceand search “elasticity” to locatecom_price_elasticity. - [04:31 - 06:45] Contract Verification: Open the Data Contract drawer; verify the 99.4 quality score and 1-hour micro-batch SLA.
- [06:46 - 08:45] Lakehouse Inspection: Preview partitioned BigQuery rows and export masked CSV samples for local modeling.
Course 02: Multi-Touch Attribution with Copilot
- [00:00 - 02:30] Context Loading: Open Group Copilot in the Cockpit; feed multi-channel advertising campaign logs.
- [02:31 - 05:40] Attribution Reasoning: Copilot evaluates
mkt_mta_touchpoints_hourlyusing Shapley Values across Google, Meta, and TikTok. - [05:41 - 08:50] Drift Correction: Prompt Copilot to adjust misattributed branded search traffic back to generic campaigns.
- [08:51 - 11:20] Prompt Standardization: Save the verified analysis trail as a reusable team prompt template.
Course 03: Sandbox What-If Simulation Boundaries
- [00:00 - 03:00] Baseline Initialization: Open
/dxp?tab=digital-twinwith $44.88M revenue and 22.50% margin baseline. - [03:01 - 06:20] Black Swan Input: Set commodity inflation to
+15%and ocean freight to$3,800/FEU; observe margin collapse. - [06:21 - 10:15] Hedging Optimization: Rebalance price tuning (+2.5%) and DTC ROAS (4.8x) to stabilize margins above 21.2%.
- [10:16 - 14:10] Execution Boundaries: Emphasize that the sandbox is a simulation; live pricing orders require PAM leases.
Course 04: Detect Sales Anomalies & Escalate to Human
- [00:00 - 02:45] Outlier Monitoring: Algorithmic scan in Commercial Tower detects a >3-sigma sales drop on an Amazon channel.
- [02:46 - 05:10] Multi-Factor Triangulation: Correlate inventory on-shelf rates, Buy Box win rates, and social sentiment.
- [05:11 - 07:30] Review Package Assembly: Generate standardized anomaly tickets with timestamps and financial impact estimates.
- [07:31 - 09:30] Dispatch Alert: Push ticket via AEP to the channel director’s mobile device for human resolution.
Course 05: Operations Review & Audit Memo Preparation
- [00:00 - 03:15] Cross-Platform Aggregation: Compare Google Ads reports, Shopify orders, and ERP ledger vouchers.
- [03:16 - 06:40] Currency & Timezone Alignment: Reconcile UTC vs. EST cutoff times and apply consistent central bank FX rates.
- [06:41 - 09:50] AEP Memo Compilation: Draft formal audit memo with underlying BigQuery SQL queries and evidence attachments.
- [09:51 - 12:05] Cryptographic Archival: Upload memo to SLSA L3 storage, generating a verified digital signature.
Operating Boundaries & Compliance Principles
- Simulation vs. Execution: Querying assets and running sandbox models are read-only analytical exercises. Live price adjustments require approved PAM leases.
- Human-in-the-Loop: All AI-assisted strategic analyses require independent human verification against canonical business logic.
- Privileged Authorization: Accessing sensitive cost columns or executing cross-tower dispatches requires active AEP grants.
Return to DXP Overview · All Video Tutorials · Console Tutorials · AEP Tutorials