Troubleshooting FAQs

This guide answers common troubleshooting questions for Connect and provides solutions to the most frequent issues.

Yes, our system does perform deletions in a few specific scenarios:

• GDPR Deletion Requests: When we receive a GDPR deletion request, we automatically propagate that deletion to the customer’s data in their warehouse to ensure compliance.

• Duplicate Row Cleanup: During the data loading process, we may remove duplicate rows in order to maintain data quality and consistency.

• Property Unsync via App UI: If a customer disables or unsyncs a property in the App UI, we treat this as a bulk deletion and drop the corresponding column from the dataset. This functionality is not currently available for DXA Connect customers.

These mechanisms ensure that data remains accurate, compliant, and aligned with customer configurations.

For data retention, we never apply data retention on data in our customers warehouses.

Sync every 24h by default. Intraday syncing will be available after GA (H2 2026).

• Daily volume ≈ your daily volume fire-hose.
  • Users = Count of Net New Users for last 24 hours
  • Sessions = Count of Unique User Sessions for last 24 hours
  • Pageviews = Count of Unique Pageviews for last 24 hours
• The _sync_info.inserted_row_count column gives an exact per-table delta after every run, so you can chart growth and cost. See Data Schema

All traffic is encrypted (TLS 1.2+); objects are encrypted at rest by the cloud providers.

CS retrieves the service-account key you supplied, instantiates the official BigQuery client library, and issues SQL INSERT/UPDATE statements over HTTPS. The TLS connection (TLS 1.2+) encrypts the queries in transit.

BigQuery will only load data files that already reside in Google Cloud Storage (GCS). Heap therefore copies the export files from its Amazon S3 bucket into a GCS bucket before triggering the load job.

Yes. Provide a GCS path that sits in your VPC and CS will stage the files there, keeping the entire transfer inside your GCP network. If you prefer not to manage a bucket, CS uses a GCS bucket in its VPC.

Heap calls jobs.insert through the same BigQuery client to create a “LOAD” job that ingests the staged files into your dataset.

Heap calls jobs.insert through the BigQuery client library with a load configuration that points to the staged files. BigQuery’s service tier then pulls the objects straight out of Cloud Storage across Google’s private backbone, using Application-Layer Transport Security (ALTS) for mutual authentication and encryption. The data never leaves Google’s internal network, and the objects remain encrypted at rest in GCS until the job finishes.

All HTTPS traffic is encrypted with TLS, and Google encrypts objects at rest in GCS. The internal GCS → BigQuery load runs on Google’s private backbone, also encrypted and authenticated.

Yes. We can switch the connector to your service account.

1. Create a dedicated service account for CS and enable BigQuery + Cloud Storage APIs.

2. Grant it the minimum roles:

   • BigQuery Data Editor on the target dataset
   • Storage Object Admin on the GCS staging bucket

3. Send Heap:

   • the service-account email
   • JSON key file (private key)
     GCS bucket name for staging

4. CS updates your integration to use that key; after the next sync the CS-owned service account is no longer used.

All subsequent file copies and load jobs run entirely under your credentials, giving you full audit visibility and revocation control.

GA4 delivers one nested events_YYYYMMDD table/day; every hit sits in an event_params array and device fields. CS Connect outputs flat, session- or page-level rows with > 200 DXA & DEM metrics (scroll depth, rage-click flags, JS errors, Core Web Vitals, etc.) and optional cvar/dvar columns. That relational shape avoids the UNNEST work GA4 requires and lets you retrofit new UX metrics without a schema change.

By default the tag masks or hashes fields that could hold Personal Data (e-mail, full IP, names, card numbers).

Yes. Use the Raw Data API: create a job, then download the gzip-compressed JSON/CSV files over HTTPS and load them yourself.

Not yet. Current integrations authenticate with a JSON key for a service account that you control. Google’s WIF endpoints are on our security backlog; earliest ETA is Q4 2025. (General WIF overview ↗ )

Data Connect is a Data Out solution only. It enables customers to sync data from Contentsquare (CSQ) to their data warehouse, but it does not support importing data back into CSQ.

To enrich sessions within CSQ (Data In), customers can use existing capabilities such as integrations (e.g., A/B testing tools, VoC platforms, analytics tools) and variables (custom or dynamic). This enriched data can then be included in the export via Data Connect.

Yes. Data is first replicated from our platform into Connect’s Snowflake account. From there, we use Snowflake’s secure data sharing ↗ to give customers access to the relevant tables and views within their own Snowflake account.

The Connect Snowflake account is hosted in the same region as the customer’s selected Snowflake destination.

Example: If the customer’s destination is in AWS eu-west-1 (EU Ireland), the data is replicated into a Connect Snowflake account in AWS eu-west-1, then shared with the customer.

Yes. Both products replicate data into a Snowflake region based on the customer’s selected destination. The data is then securely shared with the customer using the same underlying platform.

Yes, we share the data into the customers account via Snowflake’s secure data sharing ↗.

No, we no longer offer the Heap-hosted Snowflake service to new customers. Customers must have their own Snowflake account, which can be created by Sign Up for Snowflake ↗

No. The location always aligns with the region selected by the customer for their Snowflake destination.

Yes. Not all Snowflake regions are currently supported**.** A list of supported regions is available here:
On the Heap’s help docs ↗. If a customer needs a region not currently supported, a feature request can be submitted via Airfocus.

No. Both use the same platform architecture and data access mechanisms. The only difference is that Heap-hosted Snowflake is no longer being sold, even though some legacy customers still use it.