tunoc2c/workbuddy-experts/crm-hygiene

CRM Hygiene

Clean up and maintain CRM data quality: dedupe contacts, audit deal stages, update stale records, and normalize formats.

Workflow

1. Scan for duplicate contacts

• Match by: exact email, similar name + same company, same phone number
• Categorize dupes: obvious (same email), likely (name + company match), ambiguous (similar name only)

1. Auto-merge obvious dupes, flag ambiguous ones for review

• Obvious dupes: merge automatically, keeping the most complete record
• Ambiguous dupes: present both records side-by-side for user decision
• Merge rule: retain the record with more recent activity; combine any unique fields

1. Audit deal stages

• Flag deals with no activity in 30+ days
• Flag deals missing key fields (close date, amount, contact)
• Identify deals in wrong stage (e.g., closed-won without signature date)

1. Update outdated records

• Contacts with no touchpoint in 90+ days: mark for re-engagement
• Companies with no active deals: flag for review
• Stale notes or activity logs: archive

1. Normalize data formats

• Phone numbers: 统一国内手机号 11 位 / 座机区号-号码格式；区分国际号码加 +86
• Email addresses: lowercase, trim whitespace
• Company names: 统一大小写，去掉公司性质后缀（"有限公司"/"股份有限公司"/"合伙企业"/"个体工商户"/Co., Ltd.）做匹配，但保留原始名称字段
• Deal amounts: 统一为人民币（¥），外币明确标注币种和汇率口径
• 微信/企微 ID：去重时把"微信号"作为关键字段比对（同一客户可能有多个手机号但微信号唯一）

1. Generate cleanup report

• Duplicates found / merged / flagged
• Stale records identified
• Data quality score (before and after)
• Ongoing maintenance recommendations

Output Format

## CRM Cleanup Report
- Date: ...
- Contacts scanned: [count]
 
## Duplicates
- Obvious (auto-merged): [count]
- Likely (merged): [count]
- Ambiguous (flagged for review): [count] — see list below
 
## Deal Stage Issues
- Stale deals (30+ days no activity): [count]
- Missing key fields: [count]
- Wrong stage: [count]
 
## Stale Records
- Dormant contacts (90+ days): [count]
- Inactive companies: [count]
 
## Actions Taken
1. Merged [X] duplicate contacts
2. Normalized [X] records
3. Flagged [X] for manual review
 
## Data Quality Score
- Before: [score]/100
- After: [score]/100