wowprompt

Duplicate customer records matching and merge rules

Finds customers entered more than once with different spellings, phone formats or emails, scores each likely pair and sets clear rules for which record survives a merge. Use it before a CRM migration, a loyalty launch or a marketing send, when one buyer shows up as three customers.

A table of likely duplicate customer pairs with a match score and reason, merge rules for each field, and a prevention checklist for data entry
The output this prompt produced, unedited.

The prompt

Role

You are a data quality analyst who cleans customer lists. You would rather leave a doubtful pair apart than merge two different people.

Task

Find likely duplicate customer records in the list below and say how to merge them safely.

Inputs

  • Customer records: [paste rows: customer ID, name, phone, email, address, city, created date, last order date, total orders]
  • Known formats: [e.g. phones saved with and without +91, names in mixed case]
  • Fields that must never be lost in a merge: [e.g. GST number, loyalty points, consent to marketing]
  • System the merge happens in: [e.g. a CRM, a spreadsheet, an online store admin]

Method

  1. Normalise first: lowercase names and emails, strip spaces and punctuation, put phones in one format with the country code, shorten common address words.
  2. Compare candidate pairs on phone, email, name similarity and address.
  3. Score each pair: exact phone or email match is strong; similar name plus same city is medium; name alone is weak.
  4. Choose the surviving record: the one with the most orders, or the most recent activity when orders are equal.
  5. Write a field-by-field rule for the merge.

Output format

  1. | Pair | Record A | Record B | Matching fields | Score: high, medium, low | Merge? |
  2. Merge rules: | Field | Rule | Example |, e.g. keep the latest phone, add loyalty points together, keep the stricter marketing consent.
  3. Pairs to check by hand, with the question to settle.
  4. A spreadsheet formula or simple script approach to repeat the check monthly.
  5. Five data-entry rules that prevent new duplicates.

Guardrails

  • Never merge on name alone.
  • Family members who share a phone or address are different customers; flag such pairs, do not merge them.
  • Back up the list before any merge, and keep a log of merged IDs.
duplicate recordsdata deduplicationcustomer data cleaningcrm data qualityrecord matchingmaster data

How to run this

Copy the prompt and paste it into Gemini, then set the aspect ratio to 3:4. No reference photo is needed for this one.

The prompt is written in Markdown with each part under its own heading, which is what the copy action hands over. Change one section at a time, keeping the rest intact, so you can tell which change did the work.

Related prompts

All productivity