|
As a hi-tech marketer, you can appreciate new technologies and
their influences on various industries.
Advances in PC software and hardware in conjunction with revised
postal policies have brought significant increases in speed and
accuracy relating to the processing of address information.
Major changes in the direct mail industry have been driven by
postal reform policies, which outline how postal discounts are received
based on the preparation of data. Basically, if you are able to
save postal service time on its end, it will pass significant cost
savings on to you.
Cost savings occur when USPS can process mail via its automation
systems (mail pieces that conform to specific size specs and include
a barcode), save time in sorting (use of a ZIP+4 combines with the
Delivery Point directs mail to a house or building) and minimize
the amount of undeliverable mail (scrubbing data via Change of Address/Fast
Forward system).
While data preparation always has been a factor for obtaining
postal discounts, it now is the single most important component.
In a relatively short time, we have seen exponential increases
in computer processing power. A Windows NT-based server, equipped
with multiple processors and RAID storage devices acts as a powerful
host for an SQL database engine. While the BCP (bulk copy) utility
included with SQL is arcane in its command structure, it allows
the fast import of raw data into the SQL environment for performing
sophisticated queries and data mining activities. The PC-based hardware
and software solution of today rivals - and in some cases exceeds
- the processing power of larger-scale systems.
The postal changes combines with the PC technology advancements
have empowered modern-day information centers to edit, verify and
enhance the accuracy of their data as well as the accuracy of unduplication
before entering the mail stream.
To obtain a better understanding on the methodologies used, let's
examine the actual data components. Previously, it was not uncommon
for a service bureau to create what was called a hash code, a key
created by combining several elements of an address record: 5-digit
ZIP, house number, street name, last name and first initial.
For simple sorting purposes, the hash code was quite functional,
but for purposes of deduping or verification of duplication, it
was far less than accurate.
Using postal preprocessing software designed to apply ZIP, ZIP+4
and Delivery Point/Check Digit data (based on USPS postal tables),
a much more powerful sort key, or perhaps even tool, is inherently
created using the combined (12 digit) ZIP data in addition to inherent
data within the address record.
|
|
Analyzing the components, the ZIP defines the city, ZIP+4 defines
a block (a USPS- defined "sector") and the Delivery Point, which
is the last two digits of the house number, defines the actual building
within the block. Note that the Check Digit is a one-digit check-sum
to ensure accuracy of the data.
For consumer records, extracting the 12-digit ZIP information
combined with several characters of the last name can generate a
highly refined key. For business records, the 12-digit ZIP combined
with several characters of the company name will clearly define
a specific organization. Whether it is a business or a consumer
record, the resultant key is very accurate.
Understanding the ZIP components and creating a "ZIP-12" based
sort key allows you to perform some interesting activities with
the data. By importing the name and address information, as well
as the ZIP-12 key into a high-powered database engine, sorting can
be performed quickly on the key to interrogate for potential duplicated
within a file.
By further analyzing first names (if a consumer record) or the
contact name (if a business record), you could perform a fairly
aggressive duplicate elimination. However, this procedure cannot
replace what is done by industrial strength merge-purge software.
After all, using the ZIP-12 based key, the matching must be exact
or a duplicate will never be detected.
It doesn't take into account transpositions, misspellings, nicknames
or other data-centric quirks that are handled by the sophisticated
parsers within the merge-purge software.
What it can do is manipulate high volumes of data to prescreen
files in preparation for a merge-purge and matrix reporting (showing
how lists interact with each other), the scenario based on the ZIP-12
key is a viable one. Let's look at another usage possibility. The
scenario is a file of 2 million names that needed to be merged against
a collected suppress file of 20 million names from previous mailings.
Using the ZIP-12 key, you can match the 2 million-name input file
against the 20 million-name suppress file and only extract names
that have the potential of finding matching records in the merge-purge.
In a real-world situation, I have seen match rates that are anywhere
from 100 percent to 150 percent of the input file size. With this,
the suppress file has been effectively reduced from 20 million names
to 3 million.
The actual merge-purge, therefore, has been reduced from 22 million
to input names to only 5 million, which saves a significant amount
of time. As a hi-tech marketer, it's important to keep abreast of
the synthesis of tools and technologies that affect how we market
via direct mail.
In a frenetic world with tight deadlines, the timesaving and enhanced
level of analysis could greatly improve your mail plans.
# # #
|
