场景 074 · Codex Use Case

建立差异驱动桥接

综合实际、预算、预测、KPI 和负责人说明,排序重要变化驱动,输出可编辑桥接、勾稽、问题和引用。

DataKnowledge Work
Build a variance driver bridgeDataKnowledge Work

原文核对:2026-10-13 · 中文正文 · 英文提示词与演示

场景定位

综合实际、预算、预测、KPI 和负责人说明,排序重要变化驱动,输出可编辑桥接、勾稽、问题和引用。

难度
中级
时间跨度
30 分钟

适合用于

  • 预测对实际、预算对实际或不同预测版本比较。
  • 涉及收入、利润率、运营费用、现金或资产负债的分析。
  • 需要驱动排序、勾稽和负责人问题的团队。

相关工具

Spreadsheets

核对财务输入,计算变化,建立有公式检查的可编辑桥接。

Documents

需要时整理有来源的说明和负责人问题。

起步提示词

英文原文 · 可复制后按实际材料修改

Use $spreadsheets to explain the [period] movement between [actuals, budget, forecast, or prior forecast].

Use the attached close workbook, budget file, prior forecast, KPI dashboard, operating-expense tracker, cash view, and finance-owner notes. Build an editable variance bridge across the relevant revenue, margin, operating-expense, EBITDA, free-cash-flow, and balance-sheet lines.

Rank material drivers by impact, reconcile source breaks, draft owner questions, and cite the workbook tab, dashboard, tracker, or note behind each driver. Flag movements that are below [threshold] separately. Do not write a definitive explanation when the sources only support a hypothesis.

简介

差异桥接连接报告变化与运营驱动,结合各类输入排序因素,区分已支持的解释和待负责人确认的问题。

选择比较

明确期间和版本、重要性阈值、正负号约定及重点财务行。

  1. 上传实际、预算、预测、KPI 和说明。
  2. 定义比较、阈值和财务行。
  3. 要求可编辑桥接工作簿。
  4. 审阅来源断点、符号与排序。
  5. 解决负责人问题并准备解读。

检验解释

如果数据只证明变化、不能证明原因,应明确标为假设。

官方提示词

英文原文 · 可复制后按实际材料修改

Challenge the explanations in the variance bridge.

For each material driver, show:

- the calculated impact and sign
- the source file, tab, dashboard, tracker, or note
- whether the explanation is supported, inferred, or still open
- any reconciliation break
- the owner question needed to close the gap

Re-rank the bridge by absolute impact and keep unsupported explanations out of the executive summary.
展开英文原文对照

Give ChatGPT actuals, budget, forecasts, KPI data, thresholds, and owner notes, then ask it to rank the drivers behind material movements and produce an editable bridge with reconciliations, questions, and source citations.

Introduction

A variance bridge connects a reported movement to the operational drivers behind it. ChatGPT can reconcile actuals, budget, forecasts, KPI data, and owner notes, then rank the drivers and separate supported explanations from questions that still need an owner.

Choose the comparison

Define the periods and versions you want to compare, such as forecast to actual, budget to actual, or current forecast to prior forecast. Provide the materiality threshold, sign conventions, and the finance lines that matter for this review.

  1. Attach the actuals, budget, forecasts, KPI data, and owner notes.
  2. Define the comparison, materiality threshold, and required finance lines.
  3. Run the starter prompt and ask for an editable bridge workbook.
  4. Review source breaks, sign conventions, and driver rankings.
  5. Continue in the same chat to resolve owner questions and prepare commentary.

Challenge the explanations

Driver commentary should follow the evidence. Ask ChatGPT to label hypotheses clearly when the source files show a movement but do not establish its cause.

官方提示词

英文原文 · 可复制后按实际材料修改

Challenge the explanations in the variance bridge.

For each material driver, show:

- the calculated impact and sign
- the source file, tab, dashboard, tracker, or note
- whether the explanation is supported, inferred, or still open
- any reconciliation break
- the owner question needed to close the gap

Re-rank the bridge by absolute impact and keep unsupported explanations out of the executive summary.