Brand halo measurement in AMC: purchases of non-promoted products of the same brand

In short: Brand halo ASINs are products of the same brand as the ASINs promoted in the campaign. The attribution tables hold their own metrics for add to cart, detail page views, purchases, sales and units sold, each also split by click and view. That shows whether advertising triggers purchases of other products of the brand.

The fields

  • brand_halo_add_to_cart, brand_halo_detail_page_view, brand_halo_purchases, brand_halo_product_sales, brand_halo_units_sold.
  • Each has variants _clicks and _views that separate click and view attribution.
  • new_to_brand_brand_halo_* counts the same purchases, only for new-to-brand customers.
  • The totals …_total_* (for example new_to_brand_total_purchases) combine promoted and brand halo ASINs.

Example: halo share per campaign

SELECT
  campaign,
  SUM(purchases) AS promoted_purchases,
  SUM(brand_halo_purchases) AS halo_purchases,
  SUM(product_sales) AS promoted_sales,
  SUM(brand_halo_product_sales) AS halo_sales
FROM
  amazon_attributed_events_by_conversion_time
GROUP BY
  campaign
Promoted purchases and sales next to the halo purchases of the same campaign. The example has not been run on an AMC instance.

Our take

Halo purchases are attributed to the ad, not proven to be caused by it. For portfolio statements, for example in an agency pitch, the figure is an indication; for effect you need a comparison with non-exposed users, as the campaign lift query shows.

Sources: Amazon Ads API: amazon_attributed_events_by_conversion_time (retrieved 2026-10-09)

Keep reading: Query: campaign lift · New-to-brand analysis