1. What destination object would you expect to choose in the sync setup?

We would expect Fivetran to support campaign-level exclusion updates for both Google Ads and Microsoft Advertising. A suitable approach would be to introduce Campaign Exclusion Updates under the Google and Microsoft destinations.

2. What would one row in your source table represent? Can you share what your source data looks like?

Our source table contains exclusion data at a campaign level. A single row can represent up to 10,000 location exclusions (ZIP codes and/or cities) for a campaign.

Sample source data:

Expected Google Payload

{
  "customerId": "1234567890",
  "partialFailure": true,
  "operations": [
    {
      "create": {
        "campaign": "customers/123/campaigns/20987654321",
        "negative": true,
        "location": {
          "geoTargetConstant": "geoTargetConstants/9031985"
        }
      }
    },
    {
      "create": {
        "campaign": "customers/123/campaigns/20987654321",
        "negative": true,
        "location": {
          "geoTargetConstant": "geoTargetConstants/1014221"
        }
      }
    },
    {
      "create": {
        "campaign": "customers/123/campaigns/20987654321",
        "negative": true,
        "location": {
          "geoTargetConstant": "geoTargetConstants/1018862"
        }
      }
    }
  ]
}

Expected Microsoft Payload

{
  "accountId": 123,
  "campaignCriterions": [
    {
      "type": "NegativeCampaignCriterion",
      "campaignId": 123,
      "criterion": {
        "type": "LocationCriterion",
        "locationType": "PostalCode",
        "locationId": 78649
      }
    },
    {
      "type": "NegativeCampaignCriterion",
      "campaignId": 123,
      "criterion": {
        "type": "LocationCriterion",
        "locationType": "PostalCode",
        "locationId": 123
      }
    }
  ]
}

3. How would you expect Activations to identify each exclusion record?

Currently, our process is diff-based.

1. We ingest the ZIP codes and cities to be excluded from the feed.
2. We retrieve the current exclusions from Google Ads or Microsoft Advertising, including their corresponding location IDs.
3. We build the desired exclusion set from the feed data.
4. We compare the desired exclusion set with the current exclusion set on the advertising platform.

The synchronization logic is straightforward:

• Add exclusions that exist in the feed but not on the platform.
• Remove exclusions that exist on the platform but are no longer present in the feed.
• Retain exclusions that exist in both sets.

We see two possible implementation approaches:

Option 1 – Mirror Sync

• We always populate the table with the latest complete exclusion dataset.
• Fivetran sends the entire dataset and mirrors the destination state to match the source.

Option 2 – Incremental Sync

• Exclusions are loaded incrementally.
• Fivetran identifies records using a primary key and processes only newly added or updated records.

4. What fields would you expect to map from the warehouse?

Google Ads

• Campaign ID
• Customer ID
• Geo Location ID
• Action (create or remove)

Microsoft Advertising

• Campaign ID
• Account ID
• Location ID
• Action (create or remove)

5. What should happen when rows are added, updated, or removed from the source table?

Mirror Sync Approach

• No special handling is required.
• The destination should always be synchronized to match the latest state of the source table.

Incremental Sync Approach

• Only the most recent data should be processed.
• New and updated records should be applied based on the configured primary key and incremental logic.