�ݺ�ߣ

Merge Purge:

How to Do It. Why You Should Do It. What You will Learn
by Doing It.

True or false?

1. Merge Purge is a do-it-yourself job.

2. Merge Purge will prevent you from mailing the same
individual more than once.

3. Lists with a high percentage of unique records will
generally outperform lists with low %.

Direct mail is a great acquisition tool

For a cross-functional organization like the American Red Cross,
direct mail is the best channel for stewardship and stakeholder
support, and for Memorial Sloane-Kettering. “It's simply the most
cost-efficient method of acquiring new donors,” Walker says.
“Donors tend to be older. It's not because 30-something-year-olds
don't like to give, it's that they don't live at same address [for long]
and have less disposable income.”

Melissa Hoffmann Jan 01, 2012 Direct Marketing News Direct-mail heavy
nonprofits apply an integrated approach
http://www.dmnews.com/direct-mail-heavy-nonprofits-apply-an-integrated-
approach/article/220909/

… biggest ﬁnding of all is that direct-mail giving
still overwhelmingly brings in the majority of
fundraising revenue.

•Blackbaud 2011 donorCentrics Internet and Multichannel
Giving Benchmarking Report

•The2011 Target Analytics report covers 15.6 million donors
and more than $1.16 billion.

•Forlarge direct-marketing organizations that participated,
the majority of gifts still come in via direct mail.

http://www.fundraisingsuccessmag.com/article/direct-mail-still-rules-day-
fundraising/1?e=tgolab%40mediaresearch.org#

What is merge purge?

Process of combining two or more lists or ﬁles,

simultaneously identifying and/or combining duplicates

and eliminating unwanted records such as deadbeats

and nixies.

How do you do it?

Use a service bureau-software is great but having
people who can dig into data is VERY helpful

List broker, production manager, service bureau,
database manager, account manager/direct mail
manager all need to work together

How long does it take?

Once all lists have arrived
3-5 business days
Includes CASS, NCOA, deceased, de-dupe, coding,
splits, etc…
Order lists at least one week prior
Allow longer

Why you do it?

The purpose of the merge/purge is to provide a mailer with
the best possible list of names.

What Can You Learn?

Clean List or Not
Which outside lists overlap
Which lists have high match % to your house list
Which lists to cut

What can you earn?

Save on DMA pander matches, NCOA/CASS rejects, deceased,
prisons, intra dupes by NOT mailing them

Use m/p reports to support net names deals. Computer veriﬁcation
required.

Maximize your ROI by mailing multis as many times as you can.

What goes into the m/p-Categories/Status of Lists

Test Lists Continuations
• Brand new •5x Rule
• New to package
•All available
• New select
•Suppress Previous Usage

Re-Tests
• Marginal result
• Different terms
• New package

Types of Lists

Traditional
Other

•Donor List
•Buyer List •Compiled Lists
•Subscription list •Modeled Lists

House

•Lapsed List
•Non-donors, buyers,etc
•Suppression

Instructions/Cooperation

•Work with your service bureau
•Tell them whatʼs coming
•Inform them of changes
•Stay in touch-ideally have online access
•Resolving quantity variance (+/-10%)
•CASS-Blindly accept or review (i.e. apt bldgs)
•Do you need ﬁnder #ʼs?
•NCOA-How far back do you go?

Prior to m/p- Instructions to service bureau

List priority
I. Suppression

a) Current or active buyer/donor/subscriber
b) Best customers, major donor, VIP, board
members,other channel/medium
c) Internal Do Not Mail, Deceased, Chronic Non-
Responders, etc..

II. Lapsed if you wish to mail
III. Third party lists
IV. Within the lists-rental versus exchange?

Instructions to s/b-other

Appending-Finder #,congressional data, Ask string,
offer code, PURL's, CBSA, county,etc...

Splits, package tests, multis

Coding-structure (i.e.BG1FF121109)

Seeds-Yours

Coding Structure

BG1FF121109

BG=Bishop Gassis Sudan Relief Fund
1=Prospect, 2 for House/Existing
FF=From the Field (Package Name)
12-December
11-2011
09-Segment/List #

Do not blindly suppress lapsed records

•Hard suppress actives

•Separately code lapsed

•Matches between lapsed ﬁle & outside lists indicate prime lapsed
prospects

•Test mail the matches versus the non-matches

•Hint-the matches will win

Prior to M/P-delivery status

List Name Rcvd Qty. Expected Variance % Status
Qty.

List A 5,001 5,000 +1 100.02% OK to keep

List B 9,214 10,000 -786 92.14% OK to keep

List C 8,292 10,000 -1,708 82.92% Did correct file ship?

List D 11,960 11,000 +960 108.7% Continuation?

List E 0 25,000 -25,000 0% SB check for file?

Getting from input to useable records-Their
problems

List Name Input CASS NCOA DMA Deceased Intra Useable
Qty. Rejects Pander Drops Dupes Records

List A 5,001 0 24 41 131 5 4,800

List B 9,214 0 19 175 171 16 8,833

List C 8,292 2 28 225 47 10 7,980

List D 11,960 75 35 113 105 124 11,508

List E 22,575 2 80 279 256 102 21,856

Total 57,042 54,977

Getting from useable to mailable records-Your
problems

List Name Mailable In-House Multis NetNames % Unique % Unique
Records Suppression Coded by Rank
List List

List A 4,800 278 167 4,355 87% 2

List B 8,833 2,067 915 5,851 64% 6

List C 7,980 3,939 1,415 2,626 32% 13

List D 11,508 6,685 1,632 3,191 27% 14

List E 21,856 6,011 3,632 12,213 54% 8

Total 54,977 18,980 7,761 28,236 49.5%

Relationship Between Unique Percentage of
Records and Performance

List Name Qty. % Unique % Unique ROI ROI Rank
Rank

List A 4,355 87% 2- (2nd Highest) $0.26 14 (Lowest)

List B 5,851 64% 6 $0.99 10

List C 2,626 32% 13 $1.95 5

List D 3,191 27% 14-(Lowest) $3.63 1 (Highest)

List E 12,213 54% 8 $1.75 6

Multis

•Multibuyer-duplicate record on two or more customer lists,
indicating that the person has made
purchases/donations/subscribers/compiled from multiple lists, also
called multiple buyer.

•Multibuyers are likely to have a special propensity to
buy/donate/subscribe/engage-in-modeled-behavior

•List costs for multis stays with original list…ﬁrst drop

•If the prospect on List A is good (and of course they are because
why else would you and your broker have selected them) then a
prospect on List A & List B is better

Multis Detail

List Name Totals 2x 3x 4x 5x 6+
aka 2x+

List A 146 125 15 3 1 2

List B 742 585 117 26 7 7

List C 841 553 176 78 22 6

List D 1,064 738 217 64 32 11

List E 2,708 2,011 527 126 38 3

Total 5,501 4,012 1,052 297 100 29

Why is this 4,012+1,052 3x+ Total 4x+ Total 5x+ Total
total not 7,761? +297+100+2 (1,052+297+10 (297+100+29+ (100+29+
9+ (Rounds 0+29+(Rounds (Rounds 7-9) (Rounds 7-9)
7-9) 11= 7-9) 11= 1,489 11= 437 11= 140
5,501

Final Mail Files

 Records coded to lists
 Records coded to multis
 Records coded to package splits
 Records coded to package tests

List Name Qty. Code Mail Date

List A 4,355 BG1FF051201 05/15/12

List B 5,851 BG1FF051202 05/15/12

List C 2,626 BG1FF051203 05/15/12

List D 3,191 BG1FF051204 05/15/12

List E 12,213 BG1FF051205 05/15/12

List D 2xMultis + 1,064 BG1FF051206 06/05/12

List E 2xMultis + 2,708 BG1FF051207 06/05/12

Balance 2x Multis + 1,729 BG1FF051208 06/05/12

3xMultis +
1,052 BG1FF051209 06/26/12

White Mail BG1FF051299

Total 34,789

M/P-what did we learn?

1. Use a professional

2. Mail multis as many times as you've acquired them

3. Look for new lists but beware the list with a high %
of unique names

Q&A

Thank you for your time!
James Lawrence james@lawrencedirect.com
www.lawrencedirect.com
540-349-9278

Lawrence Direct Marketing Inc. Jim Lawrence

James Lawrence
@LawDirectMktng

�ݺ�ߣ

List Bazaar 2012 How and Why of Merge Purge

Recommended

More Related Content

Viewers also liked (7)

Recently uploaded (20)

List Bazaar 2012 How and Why of Merge Purge