Best Practices

Recommended setup

Group entities by purpose

Create separate data flows for operational data (buyers, contracts, opportunities) and analytics data (revenue, disbursements, insights). This keeps runs fast and makes troubleshooting easier.

Use Append for cross-cloud revenue

AWS, Azure, and GCP revenue entities share common fields like period and amount. Use the Append transformation to combine them into one unified revenue table instead of managing three separate sheets.

Join opportunities with contracts

Connect the Opportunities and Contracts entities using a Join transformation to calculate conversion rates from deal stage to signed contract — a key metric for marketplace sales teams.

Data refresh and scheduling

Match refresh cadence to entity type

Operational entities like buyers and opportunities can refresh daily. GCP monthly insights only update once a month — refreshing them daily wastes runs without gaining new data.

Use incremental GCP entities for daily tracking

For day-over-day GCP monitoring, use the incremental daily insights entity rather than the full daily insights. It returns only new changes, keeping your destination lean.

Performance optimization

Don't pull all analytics entities at once

Clazar has 15+ analytics entities. Only add the ones relevant to your active marketplaces. If you don't sell on GCP, skip all GCP entities entirely.

Send financial data to BigQuery

Revenue, disbursement, and usage entities can grow large over time. BigQuery handles volume better than Google Sheets and makes cross-entity SQL joins much easier.

Common pitfalls

Don't mix incremental and non-incremental GCP entities in the same destination table. Incremental entities return delta records; non-incremental entities return full snapshots. Appending both will produce duplicate rows.

Do

• Use non-incremental entities for initial historical backfills
• Reconcile AWS revenue against disbursements with a time-shifted comparison (30–60 day lag)
• Test with a manual run before scheduling any analytics entity

Don't

• Assume co-sell opportunities cover your full pipeline — they only reflect partner-referred deals
• Use the same API key across too many concurrent data flows — stagger runs to stay within rate limits
• Rely on Azure metered usage data if your listing uses flat-rate pricing