Skip to main content
TradeVulcan logoTradeVulcan
Work Mode Guides
Vulcan Intelligence

Analyze Google Ads and Meta performance through Windsor.ai

Request scoped campaign reports with verified fields, explicit dates and clear attribution, currency and data-coverage notes.

Vulcan Intelligence product illustration supporting the guide: Analyze Google Ads and Meta performance through Windsor.ai
Product illustration. Examples do not establish that a plugin or action is enabled for your workspace.

Start with the business question and account

Tell Work what decision the report should support: where spend went, which campaigns merit a closer look, or whether a recent creative test produced the expected platform results. Name the Windsor installation and select the correct connector/account. Use the same installation throughout the report.

Specify an explicit date range or a supported relative period. The provider’s timezone, reporting delay, attribution and conversion definitions can affect the numbers. A date range from one source should not silently be compared with a different period from another source.

A reliable initial audit prompt

TRY THIS PROMPT

Using the selected Windsor Google Ads account, audit the previous calendar month by campaign. Include spend, the conversion measures actually available, and cost per reported conversion where valid. State currency, date coverage, attribution qualifications and whether the rows are complete. Give recommendations only; do not change budgets or delivery.

For Meta, replace the source and identify the campaign/account in the same way. Do not demand a universal metric name if the connector exposes a different field. Work should discover the supported field identifiers and preserve their meanings.

Understand the discovery sequence

The useful path is selected accounts → relevant field discovery → data query. Field discovery is a catalog of available measures; it is not a campaign report and does not prove that rows exist for the chosen period.

Work uses the discovered fields for its scoped data query. It should not retrieve the entire provider metadata catalog just to answer a narrow spend question. For larger results, follow the available continuation or narrow the range and dimensions. A partial response must not become “there are no other campaigns.”

Keep platform performance and business performance separate

Google or Meta conversions can include actions that are not booked or completed jobs. Lead Forge’s reporting has its own campaign mapping and attribution. Ask Work to compare the sources only where that relationship and time coverage are verified.

If a join cannot be established, show platform metrics and business metrics separately. Do not invent bookings, collected revenue or return on ad spend from missing data. When currencies differ, do not total them without an explicit, sourced conversion method.

Make recommendations proportionate to the evidence

Ask for a short prioritized review: what was observed, why it might matter, which assumption needs checking and what change could be considered. Low volume, incomplete attribution or a paused period may make a strong conclusion premature.

A reported zero, a missing value and a failed query mean different things. The analysis should identify which one applies. A recommendation to raise a budget is not an approved provider action. Use the separate actions guide only after you decide what to change.

Export a report worth keeping

TRY THIS PROMPT

Turn the verified analysis into an Excel workbook with a campaign table, a definitions/source sheet and a clearly labeled recommendations sheet. Preserve the actual dates and missing values. Do not fabricate revenue or activate any campaign.

Wait for a ready file card and inspect the workbook. Exporting a report does not schedule recurring reporting, send it to a customer or publish it. Keep the original connection and measurement notes for the next comparison.

Provider references

Continue learning