RevOps Playbook: Data Hygiene, Deduplication, and Lifecycle Management Across HubSpot and Salesforce

Published by Bles Software, a custom software and AI company based in Yehud-Monoson, Israel, building web apps, AI agents and API integrations for clients in Israel, the US, the UK and the EU.

Dirty data undermines routing, forecasting, and attribution. This playbook provides a practical regimen for keeping people, accounts, and pipeline clean across HubSpot and Salesforce. You will implement deterministic deduplication, controlled enrichment, and lifecycle housekeeping that keeps your GTM motion fast and trustworthy.

We will establish identity keys, design merge policies, and set up periodic audits so duplicates and stale records are contained before they pollute dashboards or automations.

Who This Is For

RevOps teams that manage inbound capture, lead routing, and account ownership—and anyone responsible for reliable reporting and faster sales cycles.

Objectives

Deterministic identity: consistent keys across systems.
Predictable merges: documented rules for winning values.
Scheduled housekeeping: recurring jobs to remove drift.

Identity and Keys

Identity is the foundation of hygiene.

People

Primary key: email. Support aliases and secondary emails for known patterns.
External ID: generate and store a stable ID for cross-system alignment.

Companies/Accounts

Primary key: website domain (B2B). Normalize to root domain. Handle exceptions for multinational brands.
External ID: a warehouse or MDM identifier to anchor merges and references.

Deduplication Strategy

Adopt a two-tiered approach.

Deterministic Rules

Use exact matches on email for people and domain for companies. For companies with the same domain, evaluate country or region when relevant.

Manual Review Queue

For fuzzy matches (similar names, near-duplicate emails), create a queue with clear SLAs. Provide a merge playbook with value precedence and notes for auditability.

Merge Policy and Value Precedence

When merging, define which system or field wins for each category.

Compliance and consent: preserve opt-outs and most restrictive consent.
Lifecycle and status: choose a single source-of-truth; do not average.
Enrichment vs. human edits: prefer human edits unless empty.
Ownership: keep the active owner unless a territory rule dictates otherwise.

Enrichment and Freshness

Enrich to fill gaps, not to fight users.

Only overwrite if empty or if enrichment confidence is high.
Track last-updated-by and source for critical fields.
Review enrichment vendors quarterly for accuracy and coverage.

Housekeeping Cadence

Make hygiene predictable.

Weekly: review new dupes, fix high-severity errors, spot-check picklist drift.
Monthly: audit attribution coverage (primary campaign association), investigate null spikes, re-run dedupe on priority segments.
Quarterly: permission review, enrichment vendor check, and lifecycle stale-record cleanup.

Governance and Tooling

Store dedupe rules and merge policies in your RevOps repository. Implement helper utilities or workflows that apply these rules consistently. Use integration users for batch merges and log every action with before/after snapshots.

Anti-Patterns to Avoid

Pure fuzzy matching that generates false positives.
Overwriting human edits with enrichment on every sync.
One-off cleanup “sprints” with no ongoing cadence.
Merging without preserving external IDs and references.

FAQ

How do we dedupe across both systems without losing data?

Use external IDs and email/domain as anchors. Merge in the source system-of-record and replicate the merge to the other system. Maintain a mapping of old→new IDs for historical references.

What if two records have conflicting lifecycle states?

The system-of-record for lifecycle wins. If there’s no clear owner, define one and backfill. Document exceptions in your merge notes.

How do we manage duplicates introduced by imports?

Require templates that include external IDs and run pre-import validation. Fail fast on invalid values and teach teams to correct the file before retrying.

Can enrichment work against us?

Yes—if it overwrites human-corrected fields. Configure enrichment to fill empty fields first, and only overwrite when confidence scores clear a threshold.

How do we measure hygiene health?

Track duplicate rate, percentage of opportunities with PCA, null rates on key fields, and the age of stale records. Trend over time and tie improvements to SLA compliance.