Connect an existing listing export

What you need

Your existing listing GA4 property/stream and daily BigQuery export, its project ID and dataset location, plus a query credential with access to that export and query jobs. Reuse an existing working setup; a new Google project is not a HeyCrust requirement.

Steps

1. In Shopify’s listing tracking settings, confirm the intended GA4 measurement ID and matching stream Measurement Protocol secret for install events. Check Shopify’s listing tracking instructions.
2. Inspect your existing GA4 BigQuery link and stream selection. HeyCrust reads daily events_YYYYMMDD tables in analytics_<numeric property ID>; it does not read intraday tables. If a link is missing, your account administrator can follow Google’s link instructions.
3. Open the program’s Tracking area and choose Connect listing GA4 export. Enter the Cloud project ID, numeric GA4 property ID, numeric stream ID and the dataset’s actual location. The UI starts with US; replace it when your dataset uses another location.
4. Leave Shop URL event parameter as shop_url and App identity event parameter as api_key unless your actual export uses different names. api_key must be the Shopify client ID, not the numeric Partner app ID.
5. Supply the service-account JSON only when needed. A blank field reuses an app-saved or configured server credential; with neither available the query cannot run. The credential may originate in another project if it has the required dataset/job access. Keep JSON private.
6. Save, inspect validation, then Refresh listing attribution evidence. Follow an actual referral link/install and wait for its daily export before evaluating continuity.

Expected result

A validated project/property/stream/location and queryable daily export, followed by explicit matching evidence when actual listing and install rows arrive. Configuration success and production attribution readiness are distinct.

What HeyCrust matches

The reader looks for page_view on the canonical apps.shopify.com listing and shopify_app_install for the expected app identity. It uses exported session/user information and referral markers including hc_click, with merchant parameters such as shop_url and shop_id. This is evidence-based matching, not a promise of cross-device attribution.

Queries are bounded: four UTC days per batch, a 100 MB billed-byte cap per job, a 30-second job deadline and an overall two-minute/20,000-row read boundary. A limited read can require another refresh. Delayed exports and continuity are checked separately from access validation.

Connect this source through HeyCrust; do not change an unrelated analytics or billing setup. See Tracking verification.

Troubleshooting

Check the exact dataset, region, selected stream and service-account access. A wrong G-... measurement ID in a numeric field fails validation. Daily export can delay evidence; absence in an intraday table proves nothing for this reader.