Ad-report numbers disagree? Check the formulas

Supply a de-identified report, currency and reporting window to compare each metric's inputs and formulas before interpreting results.

Execution environment and usage

AI / software

Text-draft adaptation: ChatGPT / Claude / Gemini

Where to operate

Text chat accepting pasted report values; file upload is optional.

Getting started

  1. Open an eligible account at chatgpt.com / claude.ai / gemini.google.com and start a chat without connecting external accounts.
  2. Prepare an authorized de-identified export with currency, attribution window and timezone.
  3. Request the draft/check and compare it with the original material. Suggestions are not evidence of execution.

Setup and plan requirements

Task inputs: Prepare an authorized de-identified export with currency, attribution window and timezone. Use an eligible tool that can read the supplied format; verify parsing, context limits and access before relying on it.

Prompt portability limits

The written request can be adapted to a chatbot that accepts the same input. It does not transfer the original agent, plugins or computer permissions.

Usage limitations

Preserve report fields, currency, attribution window and time zone. Small tables may be pasted; verify parsing if uploading a file.This independent pass reviewed local case records, without revalidating original sources or hands-on testing.

Environment sources

適用情境 / When to use

Supply a de-identified report, currency and reporting window to compare each metric's inputs and formulas before interpreting results. Preserve report fields, currency, attribution window and time zone. Small tables may be pasted; verify parsing if uploading a file. Original case: Skaleit reports that Manus returned incorrect return-on-ad-spend figures in a test and corrected them after questioning.

步驟 / Steps

  1. Prepare an authorized de-identified export with currency, attribution window and timezone.
  2. Request revenue and spend inputs and calculate ratios using consistent definitions.
  3. Reconcile against the platform export; stop interpreting differences when inputs are incomplete.

提示詞 / Prompt

Check [specified metrics] for [period] in [ad-report copy]. Confirm currency, attribution window and timezone first. Show source fields, formulas, missing data and definition mismatches. Offer only testable hypotheses; do not change settings or budgets.

留意事項 / Cautions

  • Engagement and conversion are different metrics.
  • This exercise only reads reports; it changes no ads or budgets.

來源限制 / Evidence limits

This is a single agency test without independent replication; its generalizations about all brands are not adopted.

流程說明 / Workflow note

These steps and prompts are original, untested exercises by this site, not the author’s verbatim workflow. Reported results do not guarantee exercise outcomes.

原始工具組合 / Original tool stack

Manus in Meta Ads Manager (as described by the reviewer)

Source checked: 2026-10-19

Manus AI for Meta Ads: What It Can and Can’t Do (Tested on a Real Account)