# CRM Data Quality Review Playbook | 121 CRM > Find and correct records that are incomplete, duplicated, stale or inconsistent without hiding uncertainty. A fictional consultancy reviews duplicate contacts and missing ownership before a campaign. This playbook turns that kind of situation into an owned sequence with explicit stop points, exceptions and measures. Use only channels, data and capabilities approved for your setup. ## Data-quality review: the working sequence Illustrative playbook — names, organisations and activity are fictional. ## Capabilities explained ### Duplicate management Uses account rules and review steps to reduce or resolve repeated contact records. **How it works:** Stores or organises supported business data under configured fields, permissions and record rules. **Why a business would use it:** Keeps teams from following up with the same person through conflicting records. **Limitations:** The platform does not make imported, duplicated or incomplete data accurate without review. **Human responsibility:** A data owner must define the record model and resolve quality, ownership and access exceptions. ### Bulk actions Applies supported updates or actions to a selected group of records rather than one at a time. **How it works:** Stores or organises supported business data under configured fields, permissions and record rules. **Why a business would use it:** Saves administrative effort while retaining a review point before a large change. **Limitations:** The platform does not make imported, duplicated or incomplete data accurate without review. **Human responsibility:** A data owner must define the record model and resolve quality, ownership and access exceptions. ### Import and export Moves supported records through structured files, with mapping and validation needed before relying on the result. **How it works:** Stores or organises supported business data under configured fields, permissions and record rules. **Why a business would use it:** Supports controlled migration and operational extracts when the data is prepared and reconciled. **Limitations:** The platform does not make imported, duplicated or incomplete data accurate without review. Bulk migration requires mapping and reconciliation; do not treat a completed import as proof of accuracy or reversibility. **Human responsibility:** A data owner must define the record model and resolve quality, ownership and access exceptions. ### Custom fields Adds structured fields to supported record types so a business can capture information it actually needs. **How it works:** Stores or organises supported business data under configured fields, permissions and record rules. **Why a business would use it:** Replaces important notes and spreadsheet columns with searchable, reusable data. **Limitations:** The platform does not make imported, duplicated or incomplete data accurate without review. **Human responsibility:** A data owner must define the record model and resolve quality, ownership and access exceptions. ## Begin with a valid trigger, enough context and one owner A scheduled review or operational issue identifies a defined set of data-quality risks. Before any automated action, confirm that the event, customer record and communication basis are suitable for this playbook. ### Trigger A scheduled review or operational issue identifies a defined set of data-quality risks. ### Practical scenario A fictional consultancy reviews duplicate contacts and missing ownership before a campaign. ### Accountable owner Name the person or queue that accepts the work, monitors exceptions and can stop the sequence. ## Run the data-quality review playbook step by step Each step should leave evidence a colleague can understand. Timing and channels should follow the customer context rather than a generic high-pressure cadence. ### 1. Choose a bounded question Review one risk such as duplicate email, missing owner or invalid status. ### 2. Profile the records Count, sample and understand patterns before bulk changes. ### 3. Set correction rules Define what can be changed automatically and what needs a person. ### 4. Apply and reconcile Test a sample, complete the approved change and compare results. ### 5. Prevent recurrence Fix forms, imports, definitions or training that created the issue. ## Stop the normal route when the facts change Exceptions are part of the process, not failures to be hidden. Route them to a person with the authority and context to decide what happens next. ### Conflicting customer identities Do not merge until a person can establish the relationship safely. ### Legally required retention Apply the approved policy rather than deleting for neatness. ### Unknown value Record unknown or leave blank; never infer a fact for completeness. ## Measure whether the work became safer and clearer Use measures that reveal ownership, service quality and control. Counts of messages, tasks or workflow runs are supporting signals, not proof of customer value. ### Records failing the defined rule Track trend with a stable definition. ### Corrections with reviewed evidence Sample accuracy after changes. ### Repeat source of errors Prioritise prevention over repeated clean-up. ## Frequently asked questions ### When should the data-quality review playbook start? A scheduled review or operational issue identifies a defined set of data-quality risks. Confirm the record is eligible and that a named owner or monitored queue is available before allowing customer-facing action. ### Which channels should the data-quality review playbook use? Use only configured channels that suit the purpose, customer preference and consent position. Give each channel a distinct role and check provider availability and usage charges before launch. ### When should data-quality review automation stop? Stop when the purpose is complete, the customer responds or opts out, the data conflicts, or an exception needs judgement. In this playbook, pay particular attention to conflicting customer identities, legally required retention, unknown value. ### How should we judge whether data-quality review works? Review records failing the defined rule, corrections with reviewed evidence, repeat source of errors. Combine the numbers with sampled records and customer context; do not treat activity volume as a business outcome. ## Related CRM pages - [/crm/guides](https://121.direct/crm/guides) - [/crm/guides/customer-follow-up-automation](https://121.direct/crm/guides/customer-follow-up-automation) - [/crm/features/data-import-export](https://121.direct/crm/features/data-import-export) - [/crm/features/custom-fields](https://121.direct/crm/features/custom-fields) - [/crm/security-data](https://121.direct/crm/security-data) - [/crm/playbooks/weekly-crm-review](https://121.direct/crm/playbooks/weekly-crm-review) This Markdown alternative contains public product information only. It provides no access to CRM accounts, records, credentials or payment execution.