When a computer reseller wanted to increase catalog circulation last year, the company’s analysis group director told the circulation analyst to include the “ship to” address contacts in the mailing. But since the director had not checked the names for hygiene or the results of the merge/purge, he did not realize that the cataloger, which had recently combined two of its databases and computer systems, often used the data in the “sold to” field to create the shipping label, leaving the ship-to field for sales reps’ notes. The director learned about data reconciliation the hard way when 40,000 copies of the catalog were returned to the company as undeliverable.
When it comes to merging data from multiple sources — compiled files, co-operative databases, even multiple databases within your own organization — ignorance is not bliss. “Executives who aren’t involved with customer file management assume that merging files and producing customer counts and mailing lists takes only a few keystrokes,” says Bill Singleton, president of Algonquin, IL-based database marketing consultancy Singleton Marketing. “Sitting down with them and showing them a handful of examples is often the only way to demonstrate the seriousness of the data issues.”
The computer mailer’s disaster cited above could have been prevented if only the company had spent about $1,500 in having a service bureau conduct a merge/ purge of the database and some basic hygiene. Instead, “by blindly mailing the file,” Singleton says, “the mailer lost $40,000.”
CACTUS POWER
Singleton is a proponent of ensuring that your data meet the “CACTUS” standard:
Complete — with all name, address, and order history elements in place
Accurate — records must match reality
Consistent — in other words, the same field sizes and ranges of values are used throughout the organization (no notes from sales reps in the ship-to field, for example)
Timely — available when needed and up-to-date
Usable — accessible, via an efficient, user-friendly interface, to the people who need it
Secure — compliant with privacy policies and laws, and backed up in case of catastrophe.
Prospect files as well as house files must have these characteristics to support efficient and productive mailings and operations, Singleton says.
Some address elements can be verified in processing, particularly for consumer files. Take the following name and address:
J. Smith
123 E. Main St.
Brownsville, TN 38012
Consumer address accuracy can be checked against the U.S. Postal Service’s Delivery Sequence File (DSF) — which matches straight character to character — to validate the existence of the address, the range of the street number, and the directional indicator, if any.
Consumer files can also be matched against the USPS’s National Change of Address (NCOA) file to see if “J. Smith” has moved in the past 36 months. NCOA matching is exact matching, so incomplete address elements will pass through as nonmatches. Timeliness is especially important here because pulling files from beyond your 36-month archive will give you many records that cannot be updated, Singleton says.
“If you plan to try reactivation that far back,” Singleton advises, “you should process all of your accounts regularly. The processing is likely to cost much less than the catalogs you might mail to households that have moved. Running prospect and customer files through standardization and hygiene software to prepare them to match the DSF and NCOA files also prepares them for merging into a common database or for merge/purge processing.”
GET YOUR PRIORITIES STRAIGHT
Ranking your data sources for reliability and accuracy is critical when trying to resolve data discrepancies, says Don Buck, president of Milwaukee-based direct marketing consultancy Buck Marketing. Files from credit gatherers such as Experian and Equifax, for instance, tend to be highly reliable sources of information, Buck says. So if you have two similar names at the same address, such as “Jane Smith” from a credit gatherer and “Jayne Smith” from a subscription file, he advises going with the name from the credit file, as it’s more likely to be accurate.
Some mailers, however, also keep the price of the data sources in mind when reconciling conflicting data. Renton, WA-based computer gifts cataloger Computer Gear, for instance, “prioritizes co-op names over rental names because co-op names are typically cheaper than rental names,” says Mark Mackaman, vice president of operations. “So if a name is available on both a rental and a co-op basis, you want to rent the cheaper co-op name.
And don’t be afraid to rely on your house file as the ultimate hygiene, Buck adds: “The most common mistake is that marketers try to be overprecise and then end up getting the name wrong.”
For the most part, Chicago-based electrical components merchant Newark InOne favors the data from its house file when settling slight discrepancies among data from multiple sources. If “Jane Smith” were on the house file but “Jayne Smith” on a rental file, “we’d consult our house file to see if the recency is such that we have confidence mailing the name,” says Tony Chomicz, Newark’s market analyst director. “Otherwise, if the house file conveys that we haven’t heard from the customer in 24-36 months, we’ll rely on the rented name.”
Similarly, when Washington-based Smithsonian Catalogue rents names from cooperative databases, Susan Boghosian, director of marketing and strategic planning for the gifts merchant, instructs her service bureau, Central Islip, NY-based MBS Insight, to dedupe the outside names against Smithsonian’s house file — even though the co-ops suppress Smithsonian’s names when building their models.
Boghosian sees this as a way of compensating for data discrepancies that may ensue now that more customers are inputting their name and address information themselves when ordering online. In the days before the Internet, she notes, your call center representatives could simply verify the customer information, but with more and more orders being placed online, that’s becoming less and less of an option.
THE NAME GAME
During a stint as a circulation director for an apparel mailer, consultant Singleton once fielded a complaint from a customer who had received 119 copies of the latest catalog. “I’d flagged the woman as a permanent do-not-mail,” he recalls. “My investigation into the problem found that the woman’s name had been merged against 118 prospect records. The merge/purge householding option had been turned off for that urban zip code because it was dominated by multifamily housing.”
The householding option enables you to mail to just the first name that is attached to a particular address rather than to every name at the address. In areas with a preponderance of apartment buildings, such as New York City, many mailers prefer to mail to every name at a given address. This particular customer’s complete name had replaced partial, incomplete names on the other records for that address, so she ended up receiving all the catalogs intended for the occupants of her apartment building.
This example of merge/purge gone haywire demonstrates the importance of complete and consistent address information. If the merge/purge software does not strip out blanks and justify each data field element the same way, it will never match, say, “123 EMAIN” with “123EM.”
And of course, the example, just like that of the computer reseller cited earlier, demonstrates the importance of another lesson: Never assume.
The name of our magazine isn’t the only thing that’s new. We have a new Website too, complete with glossaries, calculators, and of course, years’ worth of industry-related content.
Check it out at www.MultichannelMerchant.com