Data Quality
How Merge/Purge Reduces Direct Mail Waste
Where duplicate records come from, how merge/purge matching works, how to choose the right match level and priority rules, and how to estimate what duplicates cost you.
AVS Leads · · 3 min read
Every duplicate record in a mailing is a piece you pay to print, pay to mail and hope nobody notices. Merge/purge is the process that finds and removes them. It is one of the least glamorous and most reliably useful steps in direct mail.
Where duplicates come from
- Combining lists from different sources (a rented list plus your customer file, for example)
- The same person entered twice in a CRM with slightly different spellings
- Business records listed under a trade name and a legal name
- Households with more than one family member on file
- Old and new addresses for someone who moved
How matching works
Duplicates rarely look identical. "Robert Smith, 12 Oak St." and "Bob Smith, 12 Oak Street Apt 1" may be the same person. So records are standardized first (addresses corrected, abbreviations normalized) and then compared using match rules that tolerate common variations.
Choose the right match level
| Match level | One piece per | Use when |
|---|---|---|
| Address | Address | The offer is for the household or location, not a person |
| Household | Last name at an address | Family members shouldn't each get a piece |
| Individual | Person | The offer is personal (a membership, a subscription) |
| Business | Company, or contact within it | B2B campaigns, depending on how decisions are made |
Priority and suppression
When the same record appears in more than one file, priority rules decide which version survives. Your own customer file usually wins, because it has the most accurate information. Source codes can be kept so response is still attributed to the right list.
The same process removes records that shouldn't be mailed at all: current customers in a prospecting campaign, recent responders, internal do-not-mail requests or people mailed in the previous drop.
Estimating what duplicates cost
The arithmetic is simple: number of duplicate pieces × (print cost per piece + postage per piece). As a purely hypothetical example, if 3% of a 50,000-piece mailing were duplicates, that would be 1,500 pieces paid for twice. Your own rate depends entirely on your sources and how clean they already are.
What to ask for back
- The cleaned file, with source codes preserved
- A count of duplicates removed by source
- Suppressed records kept aside rather than deleted, so you can audit them
General information only. Postal standards and legal requirements change; confirm current rules before each campaign. Questions? Talk to a data expert.