essay

Optimal_Database_Marketing_Strategy_Development_an..._----_4_-_Database_Maintenance_and_Coding.pdf

Home >Business & Finance homework help >Marketing homework help >essay

Database Maintenance and Coding

4 While carefully reviewing a draft proposal of the marketing department’s database requirements, Keri noted that the section on database mainte- nance was incomplete. She met with her marketing manager to discuss the omissions. Keri incorrectly assumed that the Information Systems (IS) department would know which fields in the database to update as promotional campaigns were run and purchases were made. To avoid miscommunications, Keri and her boss decided to systematically review all updating and maintenance requirements. They started the systematic review by examining how the fields in an individual record should change as they are entered into the database, sent promotional materials, placed an order, made purchases, or returned products. They wanted to make sure the effectiveness of lists and ad campaigns could be tracked through codes on the individual records. Whether the customer entered the database from a response to an ad in a computer magazine, filled out a form on the Web site, or came from any other medium, Marketing wanted to be able to track and evaluate the source. Customer contact information also needed to be updated as addresses, telephone numbers, and e-mail addresses changed. After Keri documented the potential updating and maintenance requirements, she was surprised at the extent of these ongoing tasks. She also realized that without periodic and accurate database maintenance, they would not be able to maximize marketing effectiveness.

Keri remembered her past experiences with improperly maintained databases. When the database was modified in DSI’s b-to-b division, sales were lost because client data were not updated properly. Keri also recalled getting irrelevant information from marketers of consumer products. For example, she received an e-mail notifying her that the manufacturer was recalling her car. She did not own that model but had only requested infor- mation about it. She also received letters welcoming her as a new catalog customer, even though she had spent hundreds of dollars over the course of

Dronzenko04 2/26/02 6:11 PM Page 55

Drozdenko, Ronald G., and Perry D. Drake. Optimal Database Marketing : Strategy, Development, and Data Mining, SAGE Publications, 2002. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/nyulibrary-ebooks/detail.action?docID=996727. Created from nyulibrary-ebooks on 2020-06-09 19:15:03.

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

56 OPTIMAL DATABASE MARKETING

several years with the company. Keri was hoping to avoid these types of problems, caused by improper database maintenance, at DSI.

Databases must be maintained and updated regularly to maximizemarketing effectiveness and efficiency. This chapter considers aspects of databases that need to be maintained, examines the importance of maintaining database fields, and provides examples on how maintenance procedures are accomplished. Database maintenance also involves develop- ing codes for records in the file. These codes help track marketing programs and eliminate inefficiencies.

Technical aspects of the database also need to be maintained. For example, IS must routinely back up data, update programs, and run tests on computer components and operating systems. Marketing may request that new versions of programs such as statistical packages be installed as they are issued. Marketing also routinely requests that IS make copies of house files for the purpose of data enhancement by external organizations. Although these types of technical maintenance procedures are important, our discussion focuses on those particular fields of the database that need routine maintenance in order to enhance the marketing process.

Standard Database Maintenance Routines ___________________

Several standard maintenance routines must be performed to keep the names and all information on the database relevant from a promotional point of view. Although software is available for many of these routines, skilled technicians are required to maintain and operate such software. The standard database maintenance routines discussed in this chapter include

♦ Deduping the customer file ♦ Householding the customer file ♦ Purging old customer records ♦ Changing contact information (address, phone number, e-mail) ♦ Standardizing addresses ♦ Removing names on databases at consumer’s request ♦ Identifying customers with match coding

In addition, we discuss the following procedures:

♦ Merge/purge processing ♦ Coding source and promotional offers ♦ Adding decoy records (salting the customer file) ♦ Identifying credit risks and frauds ♦ Field updating rules

Dronzenko04 2/26/02 6:11 PM Page 56

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

♦ Reporting summary and aggregate level information ♦ Database storage and security ♦ Database maintenance schedules ♦ Technical aspects of database maintenance

Deduping the Customer File

Occasionally, customers receive duplicates of a mailing. For example, they might receive two of the same offer for a credit card. If you review the name and addresses on the duplicate offers, there is most likely a difference. These duplicate mailings are an unnecessary expense for the marketer. Therefore, you have to dedupe your customer file.

Duplicate names on a marketing database can occur for a number of reasons:

♦ The customer’s name and/or address appeared different on an external list versus the house file

♦ For a particular order, the customer used a different mailing address ♦ The direct marketer confused the bill to/ship to addresses in the system

For example, if a marketer is mailing to current customers along with potential customers from three response lists of magazine subscribers, it is possible that some of the names on the lists and on the house file will be the same. Some may subscribe to several magazines. If names on these various lists and the house file are not properly deduped, the customer database will have duplicate customer records. To avoid this situation, a marketer per- forms a merge/purge process to the mailing file prior to sending the names to the letter shop. (This is discussed in more detail later in this chapter.)

Unfortunately, not all duplicate instances can be caught. As a result, duplicates will appear on the marketing database. Direct marketers should regularly clean up duplicate records on their database. The frequency is determined by the costs and savings.

Database Maintenance and Coding 57

Exhibit 4.1 Costs Associated With Duplicate Records

Total Number of Records

Percentage 50,000 100,000 1,000,000 Duplication Records Records Records

5 $1,250 $2,500 $25,000 10 $2,500 $5,000 $50,000 15 $3,750 $7,500 $75,000 20 $5,000 $10,000 $100,000

Dronzenko04 2/26/02 6:11 PM Page 57

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

As the percentage of duplicate records in the file increases, so does the associated expense. Exhibit 4.1 shows the costs of different levels of duplicates in files of various sizes, assuming the cost of mailing is $500 per thousand or $0.50 for each piece mailed.

Should marketers always attempt to remove duplicates from the file? The answer depends on a number of factors such as the number of duplicates in the file and the cost of the mailing piece itself. The cost to remove dupli- cates is normally minimal. Most vendors charge $0.002 per record processed for deduping. Because the cost to dedupe is minimal, you might decide to use multiple matching steps to remove duplicates.

Although there can be situations in which it is not profitable for a direct marketer to reduce waste, it might not be viewed as a good practice from a customer service or social point of view. Should the direct marketer send a mailing with a known level of duplicates that will end up in landfills or recycling centers even though they have the capacity to remove them? Will the customer become upset at receiving multiple copies of the same catalog over and over again?

If you do decide to clean the duplicate records off your database, you must put in place various rules that allow you to combine the account information of the dupes. You do not want to lose valuable information by simply eliminating one of the duplicate accounts.

Householding the Customer File

Similar to deduping a customer file, householding a customer file means examining the database for two different customers at the same address. When multiple contacts at one address are not desirable, as in the case of mailing a large annual catalog, direct marketers should identify multiple household accounts at a single address and thereby obtain promotional efficiencies and effectiveness. This allows them to avoid sending duplicate mailings to the same address. For example, both Mr. and Mrs. Jones are on the customer file. Their accounts will be identified as a “household” match, and one name will be promoted. Who is promoted from the household may depend on the individual characteristics of each member. For example, if the offer is for a cookbook, Mrs. Jones can be the best candidate.

In addition, when mailing out different offers to your customers for the same product in a single campaign, it is equally important to household the file. Not doing so can result in two people in the same household receiving two different offers. Consider the following real-life example:

Coauthor Perry Drake and his wife, Rhonda Knehans-Drake, each received a promotion approximately 2 years ago from a major credit card company at their home. Perry and Rhonda have been customers of this credit card company since 1986. This company was offering each of them

58 OPTIMAL DATABASE MARKETING

Dronzenko04 2/26/02 6:11 PM Page 58

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

the opportunity to sign up for a new credit card. Rhonda was offered the incentive of two free round-trip tickets to anywhere within the continental United States. Perry received no incentive. How do you think this made Perry feel, given that he had been a customer of this major credit card company just as long as his wife?

Be aware that customer servicing issues can also arise with householding. For example, consider the situation in which adult children are still living with their parents or where the home has mother-in-law quarters.

How household duplicates are identified is discussed later in this chapter.

Purging Old Customer Records

On a yearly basis, direct marketers should purge old inactive customers and customers whose addresses they can no longer confirm. Direct marketers should also strive to remove names of the deceased from their files to minimize the occurrence of negative reactions from consumers. Solicitations naming the deceased can be very disturbing to relatives. For example, a health insurance company in Florida experienced negative feedback from relatives of deceased individuals when offers were addressed to the deceased.

Purging old and undeliverable customer records off the database makes the database more efficient to use because fewer records have to be searched when going against the database for counts or selects. Examples of names to purge include

♦ Inactive for the past 3 years ♦ No promotions sent to the customer for the past 3 years ♦ Unable to confirm address for over 2 years ♦ Lack of external verification ♦ Deceased ♦ Nondeliverable

The details of purging rules vary from company to company. Some companies maintain two files: an active file and a purged/inactive file. A company can hold the purged/inactive file for the purpose of reactivation promotions. As we discuss in the next section, in some cases a physical address can get a response even if the current occupants have moved.

Changing Contact Information

Approximately 43 million people in the United States change addresses annually, according to the U.S. Postal Service (USPS). Nondeliverable

Database Maintenance and Coding 59

Dronzenko04 2/26/02 6:11 PM Page 59

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

60 OPTIMAL DATABASE MARKETING

mail (sometimes called nixies) should be removed from your database to eliminate this unnecessary expense. A service that helps to eliminate nixies is NCOA (National Change of Address) processing, which is licensed by the USPS. The source of the NCOA data is Permanent Change of Address forms filed by relocating postal customers. On average, NCOA contains approximately 108 million permanent change-of-address records, and the changes are kept on file for 3 years from the effective date of the move.

When a marketer sends a file to an NCOA service provider, addresses on the marketer’s file are compared with “old” addresses on the NCOA file. If a match is made, all available new addresses are returned to the marketer’s file. Approximately 13% of the records on the NCOA do not include a for- warding address. The USPS requires that NCOA maintenance be conducted in order to qualify for certain postal rate discounts. For more information on the exact requirements, visit the USPS small business Web site at www.usps.com/smallbiz and click on “How can I find and reach my target audience?” and then Business Mail 101.

Some services also check address changes against magazine publishing and credit agency data in addition to the USPS NCOA data. One reason for the additional address checks is that people may neglect to file a change of address form with the USPS when they move, but they may notify magazine publishers in order to continue to receive their subscriptions. Another reason is that NCOA matching is very strict.

Although external services are available for address and phone number updating, an organization also needs an internal updating system. Many direct marketers give customers the opportunity to provide information about change of address or phone numbers. This information should be updated on a regular basis. The address changes of an organization’s customers can be updated using NCOA services, but internal data sent directly by the customer can be more timely. According the DMA List Database Council/Research Department’s 19th Annual List Usage Survey, only 61% of responding companies use NCOA processing.

Names of people who have recently moved are called chadds in the industry. In general, chadds respond very well to promotions. Consider a selection of these recent chadds in your next promotion. Most direct mar- keters find that they can profitably promote all chadds on a regular basis.

The old address of a chadd is typically called a ghost record in the indus- try. The address of a ghost record is valid, but the current occupant is unknown. Some direct marketers keep the address of the ghost record along with relevant data about purchases by the prior occupants. For example, if the prior resident at some address purchased several gardening books from you, it is likely this address has a nicely landscaped yard. Perhaps the new, unknown resident at this same address would also be interested in purchasing garden books by way of a “current resident” promotion. In this way, the ghost record can also point you to the current address.

Dronzenko04 2/26/02 6:11 PM Page 60

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

Telemarketers should remove disconnected numbers from the database, as the cost of an average telemarketing effort can range between $1.50 and $2.50 per completed call effort. In addition to internal purging of these disconnected numbers as they occur, you can use outside services that will remove the disconnects from the database.

Removing undeliverable e-mail addresses, though costing substantially less than mail and telemarketing, still should be considered. E-mails that are repeatedly flagged as undeliverable should be removed from the database to help reduce standard database maintenance and processing costs.

Standardizing Addresses

Address quality is the key to any successful mailing project. By processing your list or file to meet USPS regulations, you can save money and stabilize postal rates. It also improves matching for purposes of deduping or householding the customer file.

The process of cleaning up a mailing list is known as address validation and standardization and is performed with CASS (USPS Coding Accuracy Support System)-certified software. This process checks each address on a list for compliance with the USPS preferred format and, when possible, converts the address to this format. It will also make changes to incorrect or missing address information on your list and add valuable postal information such as carrier routes and delivery point bar codes.

An actual example of an address before and after address standardiza- tion follows:

Address before standardization Address after standardization

6B 214 East 168th Street, Apt. B6 214 168 Bronx, NY 10456 Bronx, NY 10456

Again, to qualify for certain postal rate discounts, the USPS requires that address standardization be conducted on a set schedule. For more informa- tion on the exact requirements, visit the USPS Web site at www.usps.com and search for Business Mail 101 information.

Removing Names From Databases at Consumer Request

The Direct Marketing Association Inc. (DMA, found at www.the-dma.org) maintains a file of people who requested that their names be removed from mail, telephone, and e-mail databases. These consumers do not

Database Maintenance and Coding 61

Dronzenko04 2/26/02 6:11 PM Page 61

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

62 OPTIMAL DATABASE MARKETING

want unsolicited offers from direct marketers. The DMA programs called Mail Preference Service (MPS), Telephone Preference Service (TPS), and E-Mail Preference Service (E-MPS) are used by many national marketers to purge their files of these names. The DMA updates the MPS and TPS listings quarterly. Identifying people from the house file that do not want to be contacted saves the cost of a very low probability contact and also avoids negative consumer perceptions about a company and the industry. Here are some important things to remember regarding the preference services:

♦ Given the lead times in preparing for a mailing campaign, a consumer can still be contacted by a participating organization for a period of time after the receipt of a removal request.

♦ Not all direct marketers use the preference services, so other direct marketers might still contact a particular consumer.

♦ The preference services are typically “all or none.” However, there are exceptions. For example, a magazine publisher will still fulfill issues and ask for renewals even if the customer is on the preference list.

♦ When people on the preference list have responded to an offer, they can possibly receive future contacts from that organization.

Direct marketers are also required to remove names from databases when they receive direct requests from mail or phone call recipients. This is covered by law, and direct marketers must be aware of these federal and state laws regarding contacting customers and prospects. You can be fined for noncompliance with these regulations. For example, the Telephone Consumer Protection Act was passed in 1991. The Federal Commun- ications Commission’s rules and regulations implementing the act went into effect on December 20, 1992. One of the provisions of the act specifies that marketers must maintain a “do not call list” and honor any request to not be called again. A person’s name must be kept on the do not call list indefinitely. One error is allowed in a 12-month period. Subsequently, the soliciting companies are subject to penalties.

Identifying Customers With Match Coding

The importance of removing duplicate records at the same address has been discussed. Match coding helps in the process of removing duplicate records. Duplicate records of the same customer can occur in the house file for a number of reasons. For instance, names can be written in different ways in different databases that are being merged. The following is an example of different database presentations of coauthor Ron Drozdenko’s name:

Dronzenko04 2/26/02 6:11 PM Page 62

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

Database Maintenance and Coding 63

♦ Ronald Drozdenk (full first name and misspelled last name) ♦ R. G. Drosdenko (first and middle initials and misspelled last name) ♦ Dr. Ronald G. Drozdenko (title, full first name, middle initial, and last

name)

Another reason is that association lists might include a full name, whereas lists based on reply cards might have only initials, because people often fill them out in haste or there is a lack of space on the card. In addition, names are sometimes misspelled when they are transcribed into the database, and some consumers intentionally use variations of their names and addresses to track the sources of mailings they receive.

Match coding involves taking elements of the name and address to develop a unique identifier of an individual. Exhibit 4.2 is an example of match coding for a fictitious person living on a fictitious street in New York City.

The match code method will vary, based on organizational needs, and is typically derived internally. For some organizations, it is sufficient to code only by address and last name. Furthermore, some organizations might only be concerned with addresses if they know it is not a multiple-household dwelling. One catalog per household might be sufficient for some organiza- tions, but other organizations may want to target specific individuals within a household. For example, some political and professional organizations want to contact specific members of the household individually with application forms.

John Smith 111 Boardwalk New York, NY 10011

An example match code for John Smith is: 10011SIH111BJ

The match code is derived from the following components of Smith’s mailing address:

All 5 digits of 5-digit zip code = 10011 1st, 3rd, and 4th letters of surname = SIH Last 3 digits of street address = 111 1st street character = B Alpha initial of first name = J

Exhibit 4.2 Match Code Example

Dronzenko04 2/26/02 6:11 PM Page 63

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

A direct marketer can also use different levels of match code precision. A “tight” match requires that all or almost all characters of the code correspond, and a “loose” match requires fewer characters to correspond. Each method of matching has advantages and disadvantages. Tight code matching yields fewer matches but more are likely to be duplicates, whereas loose code matching yields more matches of which fewer are likely duplicates. Your specific needs will determine whether you need tight or loose match coding. For example, if you want to determine how much credit to extend to someone for a preapproved home equity loan, you want to apply tight match coding against the credit bureau’s file to ensure that you properly match the name to be promoted, or sent an offer, to the proper credit history data.

If the contact is costly relative to the expected return, the marketer should take measures to reduce duplicates. For example, some b-to-b organizations send customers and potential customers mailings with adver- tising specialty items (“gifts” such as pens, cups, calendars, key chains) that are often costly; therefore, duplicates should be avoided. A loose match code will reduce duplicates for these marketers. For less expensive contact methods, such as self-mailers to establish consumer awareness of a new retailer, reducing duplicates is less of a concern.

Merge/Purge Processing

Merge/purge is the process of deduping mailing lists associated with an outside list campaign. For example, two outside lists, with no duplicates in them, can be merged together and then purged of duplicates between the two lists prior to sending out the promotion. This happens because within each list individually, there are no duplicates, but when the lists are merged, there could be duplicates. For example, List A � R, S, T, and U, and List B � U, V, W, and X. Neither List A nor List B has duplicates, but if you merge them, the duplicate item will be U. The lists are merged, duplicates identified via matching algorithms, and finally one record will be eliminated from the mailing file prior to promotion.

As was the case with deduping the customer file mentioned earlier in this chapter, the most important reason for performing a merge/purge prior to mailing outside lists is the significant cost savings that can result from reduced printing and postage costs. In addition, it enhances compliance to industry regulations by minimizing the likelihood of promoting to consumers who have requested that the DMA remove their name from DMA members’ lists.

For example, in September 2000, coauthor Perry Drake, who is an adjunct faculty member at NYU, received a credit card promotion from a major credit card company that assumed he was a student. Perry has been a cardholder of this credit card company since 1986. How did this happen? Perry’s name must have accidentally been placed on an NYU student list. The credit card

64 OPTIMAL DATABASE MARKETING

Dronzenko04 2/26/02 6:11 PM Page 64

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

company must have then rented this list for promotion and decided not to apply merge/purge processing against their database of current cardholders. Perhaps the company felt that the percentage of matches would be so low that the savings in reduced mail quantities would not outweigh the cost of merge/purge processing. Unfortunately, they failed to realize the hidden cost of treating a long-time customer as a noncustomer and misidentifying his occupation.

Coding Source and Promotional Offers

Companies add unique codes to customer records to track the performance of the customers’ actions by various in-house marketing programs. Exhibit 4.3 is a mailing label that contains codes used by a company. Note the key, I.D., and media codes associated with the customer.

The catalog marketer uses these codes to monitor marketing programs and assist in order fulfillment and file maintenance.

A variety of marketing programs such as different mailing packages, media, offers, and pricing can all be uniquely coded on the customers’ record to deter- mine which marketing programs individuals have received. The codes can also be used in tests to identify which variation of a program has the best response.

Some organizations call this key coding, and some have both key codes and media codes. There is no standard method of coding in the industry,

Database Maintenance and Coding 65

BULK RATE U.S POSTAGE

PAID EDDIE CREW

3357231

853000922

Key Code

I.D. Code

Media Code

********************* ECRLOT ** R-004 Laura Martin 487 Old Kettle Rd Waterbury, MA 02798-3241

Exhibit 4.3 Catalog Mailing Label

Dronzenko04 2/26/02 6:11 PM Page 65

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

and there are a number of reasons why different coding systems are used. For example, to save space on a database file, some companies use complex codes to record a number of marketing programs (e.g., specific offers for new or existing products) that individuals have received over time. Other organizations add additional fields for new programs.

For maximum evaluation and control, each marketing activity should be identified in the database. For example, a company that markets CDs, videotapes, and books wants to relate each order to a specific mailing, direct response TV campaign, or magazine insert.

Organizations use other codes based on their needs. For example, a mar- keter may develop codes for a specific offer, catalog, test, payment method (credit card, check, money order, etc.), ordering method (phone, mail, Internet), list identification, and ad.

A direct marketer also wants to capture the source of new customers. The source field in the database indicates how a record originally entered the file. We have already discussed how an individual might enter a data- base (e.g., through rented lists, response to specific advertisements, cards filled out at a retailer, or request for information filled out on a Web site). The source is used to evaluate the performance of the method that acquired a record. Poor performing lists or promotions could be eliminated or modified. Lists or promotions that perform well could be used as models for future list acquisitions or marketing programs.

The cost of the source must also be considered in the evaluation of source performance. A higher-responding source may have an associated higher cost that does not yield a comparable return relative to a lower- responding, lower-cost source. Furthermore, the performance of sources should be evaluated over time. Some sources may have a high initial response rate but show low retention of customers over time, whereas other lower responding sources may retain more of their initial responders. These issues are discussed further in Chapter 12 in the section on Lifetime Value Methodologies.

Salting Files and Decoy Records

Owners of databases often include names in the file that are not real customers. This is referred to as salting the file with decoy records, and organizations do it to

♦ Check the performance and security of outside service bureaus that are responsible for mailing the promotional offer or fulfilling orders.

♦ Monitor delivery of packages for timed programs. ♦ Verify adherence to the list rental agreement (e.g., one-time or limited

use). For organizations that rent their lists to other organizations, the

66 OPTIMAL DATABASE MARKETING

Dronzenko04 2/26/02 6:11 PM Page 66

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

decoy names on the list can be used to detect if the list was used more times than specified in the agreement.

Mail monitoring can be outsourced. For example, U.S. Monitor offers a service that allows companies to monitor mail delivery in 100 cities in the United States. This service documents the delivery of mail pieces sent to the unique decoy names, allowing organizations to verify mail delivery, determine the length of delivery time to specific locations, and check for unauthorized use of lists.

Identifying Credit Risks and Frauds

National credit services such as Trans Union and The Credit Index flag or remove individuals from files who are considered poor credit risks. Trans Union’s national consumer credit information file includes public record information and accounts receivable data from national, regional, and local credit grantors. This detailed information is incorporated into a credit report that provides the consumer’s history of payment on financial obligations. If authorized under the Fair Credit Reporting Act, a direct marketer can use such data to extend credit, for example. The Credit Index is a co-op database compiled from various direct marketing companies that report customers who have not paid for products. This file is not strictly regulated by law, because it is compiled and shared by direct marketers about their own customers. However, The Credit Index does have its own restrictions to ensure there is no misuse of the data by direct marketers.

The Postal Bulletin (the official source of updates to USPS policies and procedures) also lists “high-risk” addresses from time to time. Direct- mailers should verify if they have these addresses and mark accordingly.

Field Updating Rules

Every time a promotion is sent to an individual on a house file and every time an individual responds to an offer, the record must be updated. Every house data item (field or variable) residing on a database has rules regard- ing how it should be maintained. For example, when a promotion is sent, various fields are updated on the marketing database:

♦ The customer is marked as being promoted for the specific offer. ♦ The “date of last promotion” field is updated. ♦ The “total number of promotions sent” field is updated. ♦ Any other fields derived from the core promotional fields are also

updated.

Database Maintenance and Coding 67

Dronzenko04 2/26/02 6:11 PM Page 67

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

68 OPTIMAL DATABASE MARKETING

When payment for the product arrives, the following fields are also updated:

♦ “Total dollars paid” ♦ “Total products paid” ♦ “Date of last payment” ♦ Any other fields derived from the core payment fields

Assume a promotion for a new cookbook was sent to customer Jones on May 16, 2000 and Jones ordered the cookbook on June 15, 2000 and sent full payment of $25.95 on July 5, 2000. If no other promotions were sent, orders received, or payments received on the Jones account between May 16, 2000 and July 5, 2000, how would the data elements shown in Exhibit 4.4 reflecting Jones’s status as of May 15, 2000 change as of July 15, 2000?

Exhibit 4.5 shows the updated Jones customer record.

Exhibit 4.4 Jones Customer Status as of May 15, 2000

Total Total Last Last Time Since Total Number Number of

Order Promotion Last Order in Dollars of Paid Promotions Name Address Date Date Months Paid Products Sent

Jones Main Street 2/15/00 3/25/00 3 $334.56 10 62

Exhibit 4.5 Jones Customer Status as of July 15, 2000

Total Total Last Last Time Since Total Number Number of

Order Promotion Last Order in Dollars of Paid Promotions Name Address Date Date Months Paid Products Sent

Jones Main Street 6/15/00 5/16/00 1 $360.51 11 63

Reporting Summary/Aggregate-Level Information _____________ In addition to updating individual records, summary/aggregate data for reporting purposes are also necessary for accurate decision-making purposes. These reports are used for marketing planning, evaluation, and control and can be updated and generated daily, weekly, monthly, and so on. Usually these reports are generated by third-party systems that are fed transactional data from the main marketing or fulfillment database in some manner. Examples of summary/aggregate fields include the following:

♦ Total mailings, calls, Web site hits ♦ Total purchases

Dronzenko04 2/26/02 6:11 PM Page 68

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

Database Maintenance and Coding 69

♦ Total orders ♦ Total orders by source, key, media, and so on ♦ Cost per order ♦ Sales per thousand contacts ♦ Profit per thousand contacts ♦ Average lifetime value

Summary/aggregate reports can be used for comparing lists, comparing offers, evaluating the customer lifetime value by source, and so on, that is, if the reports are easy to read and accurate. If you are considering the pur- chase of a campaign reporting software package, the three most important features you should look for are real-time data access, data accuracy, and system flexibility.

For example, IMT (Integrated Marketing Technology, found at www.imtnetwork.com) has produced a software package that provides real- time campaign reporting, projection, ROI calculations, and lifetime value analysis, to name a few. By linking into a direct marketer’s customer data- base, the IMT reporting software will roll up marketing information across a company’s enterprise and display it as you require. In addition, IMT will even assist you in establishing new data definitions as required to ensure that the reports produced by their software are as accurate as possible.

Examples of summary-level reports generated by the IMT system are shown in Exhibits 4.6 and 4.7. Exhibit 4.6 is a report summarizing the

Exhibit 4.6 IMT Campaign-Level Summary Report

Dronzenko04 2/26/02 6:11 PM Page 69

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

results of various new business campaigns by source. With this information, a direct marketer can determine which sources are the most profitable. Additional information such as lifetime value can be added to the report.

Exhibit 4.7 displays renewal information by source, expire date, and prior source for a magazine publisher. Additional calculations, views, or data can be added, depending on user needs. In addition, graphical repre- sentations of the data can also be generated.

Remember, you will base critical marketing decisions on this data. You have spent the money testing new campaigns and sources, so make sure you get the most out of reading the results properly by implementing reporting software appropriate for your particular situation.

70 OPTIMAL DATABASE MARKETING

Exhibit 4.7 IMT Renewal Summary Report

Database Storage and Security ______________________________ The survival of direct marketing organizations is dependent on the database. Therefore, it is essential to protect it. Database files should be backed up regularly. Backup files should be stored in a safe location (often off site) to protect them from possible damage such as fire and water. Organizations should also limit the number of personnel who are authorized to permanently modify the database. In addition, safeguards are placed in database software to reduce the chances of permanently removing data inadvertently. Encryption techniques are used to protect files from

Dronzenko04 2/26/02 6:11 PM Page 70

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

unauthorized use by internal or external sources. These encryption techniques make a file useless if unauthorized personnel access it.

Internet security is also an important concern. Organizations must have policies and procedures in place to protect and maintain the integrity of

Database Maintenance and Coding 71

Ross-Simons’ Security Statement Ross-Simons understands your concerns regarding Internet security. The Ross-Simons secure soft- ware encrypts all of your personal information including credit card number, name, and address, so that it cannot be read as the information travels over the Internet.

Ross-Simons has never had an Internet customer file a fraud complaint! The Internet is as secure as using your credit card in a store or over the phone. However, in the event of any unauthorized use of your credit card, most banks either cover all charges that result from the unauthorized use, or may limit your liability to just $50.00. Here is how the Secure Server Works

1. Starting Your Order Every order at www.ross-simons.com is secure. All data submitted via the checkout form is encrypted using SSL (Secure Sockets Layer) encryption.

2. Making a Positive ID After you’ve selected items to purchase, you’ll be prompted to continue your order by clicking on “Place Your Order” or “Place Order.” At this point, your browser looks for a “site certificate” to confirm that it is communicating with www.ross-simons.com. Once a positive ID is made, your order continues.

3. Entering “Secure Mode” At this point, your computer and our server begin communicating in secure mode, sending data back and forth in an encrypted format. We are ready to ask you for your credit card information. The encryption features of browser software are exceptionally powerful and ensure the security of this data as it is transmitted to our server.

4. Your Order Is Complete What Happens Next? The credit card numbers are placed in the SQL database on the Ross-Simons web server. A user must have login privileges on the web server (requiring a password) to be able to access the database, which also requires a password. Order information, including the credit card numbers, is retrieved by the staff at Ross-Simons through a web interface. This interface challenges the user for a password twice, and then “expires” the session after ten minutes, requiring the user to login again. All communications during these sessions are encrypted again using SSL. No credit card numbers are ever sent via email, due to the fact that email is not secure.

How can you tell if you’re in Secure Mode? The standard, unsecured URL address begins “http://”. When you enter secure mode, the beginning of the URL address will change to “https://”; the “s” stands for secure. Most browsers in secure mode also display a blue line along the top of the browser window. Netscape Navigator (versions 3.0 and earlier) display a broken key symbol in the lower left corner of your browser window when you are not viewing a secure page of a site. This key will become solid when you enter secure mode. Netscape Communicator 4.0 uses a padlock, in the same location as the key, to indicate if you are in standard or secure mode. The padlock is open in standard mode and closed in secure mode. In Microsoft’s Internet Explorer, you’ll see a padlock symbol at the bottom of your browser window when the browser is in secure mode.

Exhibit 4.8 Ross-Simons’s Security Statement

Source: www.ross-simons.com/custom/security.html

Dronzenko04 2/26/02 6:11 PM Page 71

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

customer data. Customers are more likely to purchase online if they feel transactions are secure. Exhibit 4.8 is an example of how an online business, Ross-Simons, is attempting to reassure consumers that transactions are secure. This company sells fine jewelry, tableware, gifts, and collectibles, with many items costing thousands of dollars; therefore, making customers feel safe is important. Many customers are also concerned about personal data being accessed and used; privacy issues are discussed in Chapter 17.

72 OPTIMAL DATABASE MARKETING

Database Maintenance Schedules ___________________________ There is a cost to database maintenance. Merging and purging names, flag- ging credit risks, updating file fields, backing up files, and so on require processing and human resource time. Therefore, maintenance should be performed at intervals that minimize maintenance costs while optimizing marketing efficiency and effectiveness.

Aging data on the database have a deteriorating impact on marketing effectiveness and efficiency. All house data elements should be routinely updated. How often direct marketers update the customer promotion, order, and payment fields residing on a marketing database will depend on how often they select names from the database for promotion. Major direct marketers with weekly campaigns update the customer data at least once per week. Others, having only one or two campaigns per month, require a monthly update.

You should update enhancement data a minimum of once per year. Lifestyle and demographic data about a customer can change quickly. Over the course of a year, a person can become a new parent, take on a new and higher paying job, take up golf, or decide to become involved with a char- itable organization. Depending on how valuable you find this type of data in helping you better understand your customers’ needs and target them with promotions, you might decide to update it twice per year. Your needs define the frequency of your updates.

Exhibit 4.9 gives some guidelines commonly used in the direct mail industry for the maintenance schedule of other routines.

Exhibit 4.9 Standard Maintenance Schedules

Process Common Schedule

NCOA processing Monthly or quarterly for large mailers; two times per year for others

Address standardization & customer file deduping As needed for legal and postal compliance

Householding of the customer file Annually for clean-up of duplicate records

Application of DMA do-not-promotes As needed or quarterly. The MPS and TPS are updated quarterly by the DMA

Dronzenko04 2/26/02 6:11 PM Page 72

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

The best time to establish database maintenance rules and procedures is when the database is developed. The data fields that are to be updated when promotions are sent, orders are placed, or payments are made should be determined during the database planning stage. However, it is often impos- sible to anticipate all the variables that might be useful for data analysis. In some situations, the main database will not contain all the updated variables. Other files are extracted from the main database for analysis purposes, and these files will contain new variables that the analyst feels might be useful for decision making.

Depending on the database system, various techniques are used to maintain and update files. Most relational databases are manipulated with structured query language (SQL) programs. SQL is flexible and relatively easy to use. Older database systems use COBOL programming language. Despite the fact that this program is decades old, COBOL programming language is still commonly used by database marketers. Some newer applications programs such as data miners use graphical interfaces that make data manipulation and analysis easier.

Database Maintenance and Coding 73

____________ Some Technical Aspects of Database Maintenance

________________________________________ Chapter Summary Customer databases are constantly changing in marketing-focused organiza- tions. The database needs to be updated as new customers come onto the file, existing customers change addresses, promotions are delivered, orders are placed, and payments are made. Marketers also have to remove names of individuals from lists who do not wish to be contacted. Improperly main- tained databases can reduce marketing efficiencies and effectiveness. Although there are costs associated with database maintenance, greater costs are usually incurred when they are not maintained. In addition, the effectiveness of marketing programs can be tracked by entering codes into the database. These codes tell marketers about the effectiveness of promotions, lists, media, offers, and so on. Because databases require substantial invest- ments and represent a critical asset to the organization, procedures should be in place that secure the database from accidental or intentional misuse.

________________________________________ Review Questions 1. What are some of the reasons for routinely maintaining the

database?

2. What is meant by merge/purge?

3. What are some of the codes marketers place on individual records, and how are they used?

Dronzenko04 2/26/02 6:11 PM Page 73

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .

4. Why would a marketer put decoy records on a database?

5. When and how is a database updated? Give examples of some of the fields that would be routinely updated.

6. Why is database security important to both consumers and marketers?

74 OPTIMAL DATABASE MARKETING

Dronzenko04 2/26/02 6:11 PM Page 74

C o p yr

ig h t ©

2 0 0 2 . S

A G

E P

u b lic

a tio

n s.

A ll

ri g h ts

r e se

rv e d .