Practical guidance for handling Large Data Volumes (LDV) in Salesforce: key design patterns, indexing, async processing, archiving, and operational tactics to keep your org fast and scalable.
What is Large Data Volumes (LDV) in Salesforce?
Large Data Volumes (LDV) refers to scenarios where objects contain hundreds of thousands to billions of records. At this scale, normal operations—queries, reports, list views, triggers, sharing recalculations, and data loads—can slow down or hit governor limits. LDV is contextual: an object becomes LDV when default configuration and standard optimizations no longer deliver acceptable performance.
Why LDV Matters
When LDV issues appear, users and integrations experience slow responses, failed data loads, record-locking problems, long-running Apex jobs, and increased storage costs. Addressing LDV early with the right architecture prevents outages, maintains SLAs, and reduces long-term maintenance burden.
Top 10 Best Practices
- Choose the right data model — Keep relationships minimal for high-volume objects. Use master-detail where appropriate, archive unused lookups, and avoid many-to-many patterns for hot tables.
- Index fields and write selective SOQL — Use indexed fields (ID, owner, audit fields, external IDs) and avoid non-selective operators (NOT, !=, CONTAINS, LIKE ‘%’). Request custom indexes from Salesforce when needed.
- Use skinny tables — For frequently-read fields on very large objects, skinny tables reduce I/O by storing only selected columns for faster queries.
- Archive or purge old data — Move historical or low-value records out of standard objects to lower storage, speed up reports, and shorten sharing recalculations.
- Use asynchronous processing — Offload heavy work to Batch Apex, Queueable, Platform Events, or external jobs to avoid synchronous governor limits.
- Optimize triggers — Adopt one-trigger-per-object patterns, bulkify logic, and avoid SOQL/DML in loops to prevent cascade failures at scale.
- Use Big Objects for historical/event data — Big Objects scale to hundreds of millions/billions of rows and are ideal for logs, audit trails, or telemetry.
- Leverage external systems — Virtualize large datasets that don’t require local storage. Reference them on demand instead of synchronizing everything into Salesforce.
- Optimize reports and list views — Design bounded filters and use indexed fields so analytics don’t trigger full table scans or dashboard failures.
- Avoid data skew — Distribute ownership and parent relationships to prevent record-locking, UNABLE_TO_LOCK_ROW errors, and sharing recalculation hotspots.
Practical checklist
- Run query/selectivity audits on large objects.
- Identify candidate fields for indexing or skinny tables.
- Plan archiving retention (what stays in Salesforce vs external store).
- Convert synchronous processes to async jobs where possible.
- Review triggers and bulk patterns across objects with >1M rows.
Conclusion — Why this matters for admins, developers, and business users
Handling LDV is not a single feature flip — it requires architecture, operational discipline, and collaboration between admins, developers, and stakeholders. Admins guide retention and reporting rules; developers implement selective queries, async patterns, and efficient triggers; business users get reliable performance and timely insights. Investing in LDV best practices preserves system health, keeps costs manageable, and enables growth without disruption.