essay

Optimal_Database_Marketing_Strategy_Development_an..._----_5_-_Basic_Database_Technology_Organizational_Considerations_and_Databa....pdf

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

Basic Database Technology, Organizational

Considerations, and Database Planning

5 Keri Lee understood that the development of DSI’s consumer database involved two processes. One process was developing the marketing specifications. The other process was technical development. Technical development involved determining the computer hardware and software that would be required for the marketing specifications. Because Keri grad- uated with a minor in management information systems (MIS) and had a good working knowledge of DSI’s b-to-b database, DSI’s marketing manager decided that Keri would be a representative from Marketing on the database development team. The development team was cross- functional and was comprised of people from MIS, Marketing, Finance, Accounting, Human Resources, and Operations. DSI’s CEO was also a member of the team.

At the first meeting of the team, Marketing submitted their database specifications. These specifications included the potential size of the database in terms of records and fields, the data fields initially needed, updating requirements, and formulas for new variables derived from the data fields. In addition, Marketing wanted the database to be flexible so that new fields could be added easily and routine information could be extracted from a product manager’s workstation quickly for monitoring and controlling marketing programs. Keri’s supervisor, who is experienced in database analysis, also insisted that the database must be designed to allow easy access to data for transporting into statistical analysis programs.

The other issue that was critical to Marketing was Internet commerce. Because DSI’s target market is computer savvy, Marketing wanted an integrated, interactive database in which customer data could be tracked across mail, phone, and Internet transactions. Marketing felt there was great potential for upgrading, upselling, and cross-selling with an Internet database. Customers could be prompted on new products and upgrades

Drozdenko05 2/26/02 6:12 PM Page 75

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 19:15:14.

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

76 OPTIMAL DATABASE MARKETING

through e-mail or timed messages in DSI’s programs that would refer them to the company’s Web site.

MIS immediately recognized that existing technological resources at DSI were inadequate to meet these requirements and that new hardware and software would have to be purchased. MIS would study Marketing’s requirements and develop a proposal for technical development. Because new resources would be needed, the VP of Finance would also be involved with proposal development. Because resources were limited, Marketing might need to consider accepting some compromises. Furthermore, DSI’s corporate goal was to develop a database that was integrated across all business functions. The other departments also would submit their database requirements. Because DSI had a strong customer focus in their mission statement, the CEO indicated that Marketing’s specifications would have a high priority.

The meeting ended at 6:45 p.m. As Keri left the building for the day, she realized that the task of developing the database was substantial. There might be tension among the departments, and she anticipated many more long meetings. She already had some ideas on what types of hardware and software might be able to meet Marketing’s requirements. Keri knew that a relational database offered the needed flexibility, but she was not sure which software package would be ideal for the specifications of all the departments. She was anxious to see MIS’s proposal. The development of the database would be the start of the career challenge Keri was seeking.

In most organizations, Marketing is not responsible for establishing thetechnical specifications of the database or for maintaining the database. However, marketing personnel should have a basic understanding of data- base technology. This knowledge is important because it helps marketers contribute to database system design from an applications perspective. Knowing general concepts of database technology provides the marketer with insights to opportunities and limitations of certain database systems. If marketers understand these opportunities and limitations, they will be better able to establish marketing specifications for the database that are reasonable and will maximize effectiveness. In addition, marketers who understand the basics of database technology are better able to interact with MIS professionals.

In some smaller organizations, marketing personnel may have primary responsibility for establishing the customer database. The technological considerations are less of a problem when organizations plan to build a relatively small database housed on a PC. If marketing personnel have a good working understanding of PCs, they should be able to establish a simple relational database using one of a number of user-friendly database

Drozdenko05 2/26/02 6:12 PM Page 76

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

Modern marketing databases require computer hardware and software. Hardware is the physical equipment that holds the database. It includes processors, storage devices, input and output devices, and components that link devices together in networks. The term server refers to hardware that links devices in a network, such as PCs, storage devices, Internet communi- cation devices, and printers. Exhibit 5.1 is a simplified diagram of a computer network.

Networks can be limited to one location or link several locations. In addition, servers can link the company with customers and employees through the Internet. The challenge for many companies is to develop

Technology, Organization, and Planning 77

________________________ Computer Hardware and Software

programs. As the complexity of the database increases, however, people with more technical knowledge of database development should be consulted.

In this chapter, we discuss computer software and hardware available on the market today, options for database organization, query languages available for pulling customers and customer information from the data- base, and other issues that must be addressed to ensure success. In addition, we discuss in detail how to select and manage a database provider if you decide to outsource the building of your customer database.

Server Server

Product Manager

Analyst

Marketing Manager

Storage Output

Devices

Processor

Internet

Customers

Employees

Exhibit 5.1 Simplified Diagram of a Computer Network

Drozdenko05 2/26/02 6:12 PM Page 77

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

78 OPTIMAL DATABASE MARKETING

database systems that have the capability to handle the transition to Internet commerce.

Software comprises the programs of instructions to the hardware that allow data to be filed, sorted, analyzed, input, and output. Hardware and software have to be compatible. Therefore, it is advantageous to develop databases as a system rather than purchase hardware and software independently.

IBM and other companies offer preconfigured systems for specific purposes. These preconfigured systems can be valuable for companies that are concerned about compatibility problems. For example, IBM combines S/390 server technology with database, data mining, and operating system software. IBM also provides technical support on the system. This system will facilitate the database marketer’s ability to store and analyze customer data (IBM, 2001).

In contrast, some companies prefer to configure their own systems. Although this takes more time and more expertise, it has advantages. The company can select hardware and software from a number of vendors to optimize the system for their specific needs. Recently, hardware and soft- ware manufacturers have been more cooperative in developing products that are compatible. This means that it is easier to find optimal components for database systems that are compatible.

Database Hardware

Database hardware currently can be divided into three categories: (1) main- frame, for large databases, (2) midrange, for medium to large databases, and (3) PC, for small to medium databases.

The specifications of these categories change as new generations of com- puter technology are developed. Computers have been able to handle larger and larger databases. Powerful midrange computers can handle databases that previously required mainframes, and PCs can handle databases that pre- viously required midrange computers. However, as technology has advanced, so has the need for more extensive databases. Internet commerce and the global linking of databases require more powerful technology. Consequently, the predicted demise of the mainframe computer never occurred.

Mainframe computers, such as the IBM 390 series, can hold millions of customer records and are still important for large database marketers. Physically, newer mainframe computers have been reduced in size due to technological innovations. Older mainframes can span several hundred square feet and weigh more than 30,000 pounds. Newer mainframes can be less than 60 square feet and weigh a few thousand pounds. The newer mainframes also cost less, are more energy efficient, and generally are faster. Mainframe computers are also called enterprise servers and superservers because they are used for large client/server networks and high-volume Internet commerce sites. First Union Corporation, for example, uses an

Drozdenko05 2/26/02 6:12 PM Page 78

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

Amdahl mainframe system. First Union provides financial services to more than 16 million customers throughout the East Coast and other parts of the United States. First Union needed a highly reliable system for its customers who rely on the bank for cash from its ATMs. In addition, customers want to be able to do banking any time of the day, either in person or over the Internet. The requirements of First Union’s system warranted the use of powerful mainframes (“First Union,” 2001).

Midrange Computers

Midrange computers, also called minicomputers, are used by many organizations that don’t need the computing power of mainframes. The IBM AS/400 is an example of a midrange computer. Midrange computers have also been used as network servers to manage Internet Web sites and internal and external networks. In some applications, midrange computers interact with mainframes.

Tiffany & Company, noted for its jewelry and home décor items, is an example of a company that uses several midrange computers. Tiffany has retail stores around the world and a catalog business. A primary objective was to allow managers to easily access data in order to understand and respond to changing customer needs. The system is integrated across functional business areas and provides a unified decision-support system (“Tiffany & Company,” 2001).

PCs

PCs, also called microcomputers, have evidenced a tremendous advance in technology in the past few years. PCs now have power that exceeds the mainframes of previous computer generations. The range of PCs is quite wide, varying from desktop units to notebooks to handheld units. A number of companies manufacture PCs, including Dell, IBM, Gateway, Compaq, and Hewlett Packard. Many companies now produce customized computers for individual customers. Interactive Web sites (e.g., www.dell.com) allow customers to configure and order a computer. Customers can select compo- nents such as processors, memory, disk storage, modems, sound cards, video boards, monitors, and keyboards. Software such as operating systems and business application programs can also be selected and loaded into the computer.

PCs can hold more than a million database records at this time, and their capacity continues to grow. They are often linked with mainframes and midrange computers in a network. There are substantial advantages to using network systems for marketing databases. Many midsized and large

Technology, Organization, and Planning 79

Drozdenko05 2/26/02 6:12 PM Page 79

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

80 OPTIMAL DATABASE MARKETING

companies house the database on mainframe or midrange computers and use PCs to access data for manipulation and analysis. Product managers and other marketers can analyze and track the performance of marketing programs from their PCs.

Databases housed on PCs are very common. For example, single location retailers of all types develop customer databases on PCs. In addition to handling accounting functions, suites of programs for retailers also allow the small business owner to analyze customer databases and link the databases to other software used for mailings. Recently, small businesses have been able to establish e-commerce sites with PC-based software. One of the concerns is the ability of the database software to be upgraded as the com- pany grows. Therefore, more vendors of small business database software are making their programs compatible with other software and hardware.

Hardware Decision Factors

When hardware from a particular vendor is evaluated, a number of factors must be considered. These include cost, performance in terms of speed and capacity, reliability, compatibility, and service support. In addition, it is desirable to have hardware that can be expanded and upgraded easily and can handle the demands of a wide range of users and their processing requirements.

MIS professionals have to consider both current and future needs of the organization. Lower-cost hardware may have limitations for future expan- sion as the organization’s processing needs grow. Changing hardware and software is expensive and disruptive to the organization, so expandability is a high priority for many companies. Some companies with larger data- bases have faced (and are still facing) a long conversion process as they upgrade older mainframe systems (legacy systems). This conversion process can take years and cost millions of dollars.

Database Software

To run a database, several software programs are required. Operating sys- tems software, like Windows NT and UNIX, provide a user interface and manage computer resources, tasks, and files. Other programs such as peripheral controllers, virus protection, and data security programs are also necessary to support the information system.

Database software controls the storage, organization, analysis, and inter- faces between people and databases. Smaller organizations that run their businesses on PCs can use general-purpose relational databases like Microsoft Access (see Exhibit 5.2) and Lotus Approach.

Drozdenko05 2/26/02 6:12 PM Page 80

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

These programs are easy to use and are adaptable to various business applications. Specialized database software such as ACT! (see Exhibit 5.3) and GoldMine allow small- to medium-sized businesses to manage customer relationships. Some of these customer contact programs allow multiple users and interfacing to Web sites for e-commerce.

Larger organizations that have more complex databases and run on mainframe or midrange-based systems often have database administrators and support staff that manage the database for the entire organization. These organizations have database management systems (DBMS) (see Exhibit 5.4), which are a set of computer programs that control the development, maintenance, and use of the database.

Oracle, SAP, Sybase, and DB2 are database programs used by larger organizations. Oracle, for example, offers database programs for a wide range of business applications. These programs are adaptable to many different business functions and industries. Oracle Marketing is a suite of applications designed to aid in the planning, execution, and analysis of marketing campaigns. Oracle also offers a suite of programs for Internet commerce applications. Companies like Amazon, eBay, and E*TRADE are reported to use Oracle software.

Because there are several producers of reliable and adaptable database programs, there are fewer reasons for companies to write database

Technology, Organization, and Planning 81

Exhibit 5.2 A Screen From Microsoft Access

Drozdenko05 2/26/02 6:12 PM Page 81

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

82 OPTIMAL DATABASE MARKETING

programs from scratch. Developing database software for large business applications is a complex and time-consuming process. The major database producers possess the expertise and support staff to continually improve and expand the database applications. In addition, they are constantly obtaining feedback from the companies that are using their software. Therefore, it would be difficult for a single organization to match this level of database development expertise.

It should be noted that many direct marketing organizations have developed their own database programs over the course of many years. This was a common practice before the widespread availability of commercial database programs. These companies have been constantly upgrading their databases using in-house programmers. As the business evolves, the company modifies the database programs or adds other programs to increase access or analysis capabilities. Therefore, these companies have a large investment in their current database, and regardless of the advantages of new technologies, it is not beneficial for them to change database systems. Some of these companies will upgrade in small steps to reduce the impact on daily operations.

Exhibit 5.3 A Screen From ACT!

Drozdenko05 2/26/02 6:12 PM Page 82

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

Technology, Organization, and Planning 83

Customer Service

Sales

Promotion

D B M S

Report

Analyze

Sort

Customer Data Transactions

Calls, Etc.

Exhibit 5.4 Database Management System (DBMS)

____________________________ Database System Organization

Database systems must be organized in some manner to allow data to be retrieved. These systems can be categorized as either structured or relational, depending on how the data are stored. Data in structured databases come from a single source, whereas data from relational databases come from multiple sources called tables. Recently developed marketing databases are usually relational. However, some larger marketing databases that were developed in the past on mainframes still have a structured design.

Structured Databases

Structured databases can be further classified as flat, hierarchical, or net- work. In flat files, data are accessed in a sequential manner, and it is difficult to add new data fields. For example, to access the customer John Smith in a flat file that is organized alphabetically, you have to pass all customers with names ending with A, B, C, D, and so on. Flat files can have fixed or vari- able formats. With fixed formats, all customer records are of the same length, and data input to fields is limited by the space allocated. For example, if the

Drozdenko05 2/26/02 6:12 PM Page 83

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

field for a last name is fixed at 15 spaces, all last names longer than 15 char- acters will have to be truncated or somehow abbreviated. The number of fields are also limited in a fixed format database. For some database applica- tions, it becomes impractical to use the fixed format. If you expect a large variation in purchase activity, a variable file format is more efficient than a fixed format. Using the fixed format, sufficient space needs to be allocated on each record to anticipate the largest number of transactions possible. With the variable format, the size of the field adjusts according to the number of transactions. Indexed flat files are linked together by a unique customer code. The index reduces the time needed to search through the file.

Hierarchical databases use a tree structure for searching data. Each data element is related to only one element above it, so the relationships between the data items are called one-to-many. Although hierarchical databases use a predetermined path, access to data is quicker, because the data items are linked in a meaningful way.

Network structures still use predetermined paths to access data. However, they allow many-to-many relationships. That is, there may be two or more paths to access a data item. In the network structure in Exhibit 5.5, the data item “price” can be accessed through two paths. In a hierar- chical structure, there would be only one path to access “price.”

84 OPTIMAL DATABASE MARKETING

Sales Rep

Customer

Order Number

Price

Part Number

Exhibit 5.5 Example of Network Structure Paths

Drozdenko05 2/26/02 6:12 PM Page 84

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

Relational Databases

Relational databases have become the database of choice for most organ- izations that are developing new databases systems. Relational database programs are available for PCs, midrange, and mainframe computers. No predetermined structural relationships between data items are needed to access information. Data are organized into tables that can interact with each other. For integrity of data, similar data items in different tables are linked to each other. Changes in one item must be accompanied by changes in the corresponding item(s). Relational databases offer a high degree of flexibility compared to structured databases.

Multidimensional databases are a variation of relational databases. Instead of using tables that are two-dimensional, multidimensional databases are structured on many dimensions. This is useful for some analytic procedures in which dimensions could represent product category, region, distribution channel, and time.

Comparison of Structured and Relational Databases

Although relational databases have some distinct advantages for flexibil- ity, there are still reasons to consider structured databases. In the past and for some current larger systems, run time to update relational databases would be excessive compared to structured flat file architecture. However, as processing speed continues to increase and the cost of computer mem- ory decreases, the advantages of structured databases will diminish. Exhibit 5.6 summarizes when to use structured or relational databases.

Technology, Organization, and Planning 85

_________ Structured Query Language (SQL) and Data Analysis

Data analysis is an essential element of strategic database marketing. Some software programs make the extraction of names for promo- tion and analysis easier and more reliable. SQL is a widely used

Exhibit 5.6 Application of Structured Versus Relational Database

Structured Relational

Data relationships form a natural order or hierarchy Flexibility to examine complex relationships

Data items do not change much over time Ability to look at the big picture

Fine-tuning important Storage efficiency important

Computer power conservation important Development speed important

Source: Adapted from Jackson & Wang, from Strategic Database Marketing (1997, p. 128).

Drozdenko05 2/26/02 6:12 PM Page 85

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

programming language for relational databases. In contrast to third- generation programming languages like COBOL, SQL is relatively easy to use. It allows data to be extracted from large databases for promotion and analysis. For example, with just a few SQL commands, customers who have been sent a specific promotion can be easily selected. Once these customers are extracted from the database, additional analysis can be performed.

Statistical analyses are commonly performed on database files. Because of the availability of reliable statistical analysis programs, few organizations write their own statistical programs. SAS is an example of a statistical analysis program that is widely used by database marketers. SAS procedures are usually performed on portions of the database, so data extraction and conditioning is necessary before an analysis can be performed. SAS is often used for modeling, which allows the marketer to determine who the best customers are, based on certain characteris- tics. In Chapters 6 to 14, we examine statistical procedures in more detail.

86 OPTIMAL DATABASE MARKETING

Organizational Considerations in Technical Database Design _____________________________

As we discussed, marketing professionals will not have primary respon- sibility in most organizations for selecting the hardware and software that will be used in a database. In most large organizations, the IS department (sometimes referred to as IT, information technology) is responsible for the technical development. Often a cross-functional development team is established to develop the database. Members of the development team should have a common goal: to develop an effect- ive and efficient database that can be used to achieve organizational objectives.

In practice, this is not an easy goal to achieve. Marketing and IS people have different perspectives and sometimes different definitions of key object- ives. As part of this cross-functional team, marketing professionals should strive to look at the database development process from the perspective of the IS professionals, and the IS professionals should strive to look at the process from the perspective of marketing professionals. Stephen Belfer (1998) pointed to the different perspectives that marketing and IS profes- sionals possess based on their backgrounds. Marketers tend to be “project” people. Marketing programs must be changed in order to respond to changing market forces. As one marketing campaign ends, another begins. In contrast, IS professionals are much more “process” oriented, because their role is to maintain and enhance the overall information technology infrastructure of the company.

Drozdenko05 2/26/02 6:12 PM Page 86

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

Individual hardware or software projects are only components of the ongoing development of the greater information system. Therefore, Marketing has to understand that IS has to consider not only Marketing’s immediate needs but also how specific hardware and software decisions affect the entire information system in the long term. In addition, because IS may be working within the constraints of existing hardware and software, it may not be feasible for them to simply purchase and install a commercially available database program.

Another difference in perspectives involves defining the characteristics of an ideal database system. IS would tend to be biased in the direction of developing systems that are technically efficient. That is, the goal is to minimize storage and processing requirements. On the other hand, Marketing sees the ideal system as flexible and readily accessible. Flexible and accessible systems tend to be less efficient.

To facilitate the development process, Marketing should clearly specify the following:

♦ Number of records and fields required Present and future needs should be considered.

♦ How and when the records and fields need to be updated The required cycle for database maintenance should be specified.

♦ How data or records will be entered or updated This should consider data from external sources such as lists and determine if customer data come from multiple sources such as phone, mail, and Internet.

♦ How and by whom the database will be accessed For example, product managers and analysts should have access to the database from their PCs. Internet and other remote access capabilities should also be considered.

♦ Data analysis requirements This should include sorting and selection capabilities, statistical analysis packages, and periodic reporting requirements.

♦ Data security considerations Specify who within the organization should have access to data. Data manipulation privileges may also need to be restricted. With Internet commerce, the issue of security becomes more complex. This must be resolved through consultation with data security experts.

When IS receives these requirements from Marketing, they will determine what software and hardware will be required. On the basis of their background and personal experience, some marketing personnel may be able to suggest or require certain software programs. For example, because certain marketers or analysts are familiar with the SAS statistical package, Marketing may want to design the analysis capabilities of the database around SAS.

Technology, Organization, and Planning 87

Drozdenko05 2/26/02 6:12 PM Page 87

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

88 OPTIMAL DATABASE MARKETING

Outsourcing: The Process to Select a Database Provider _____________________________

In Chapter 3, we discussed customer data requirements. When examining the customer data requirements, an assessment of corporate information needs is conducted. When corporate data needs relative to the customer base are assessed and the associated data sources are identified, a corpor- ation must make a decision: build and manage the database in house or outsource to an external vendor. This decision will be influenced largely by the number and complexity of data feeds required to build the database. A company with a single data feed may opt to build and manage the database development internally. A company with a dozen data feeds may opt to place the database maintenance and development in the hands of experts. In addition, if the situation is to create a marketing database for the first time, a company would be well served to hire a consultant because of the complexities and time-consuming nature of planning for a corporate data- base.

The advantages to outsourcing the building and management of the database include the following:

♦ The organization does not need to maintain the staff to administer the technological or data processing support for the marketing database.

♦ Existing IT staff can perform their functions with minimum disruption. ♦ The task of building and maintaining a database is placed with

experts. ♦ External vendors are more conversant with cutting-edge technology

for management and analysis of data than internal staff.

Because many organizations consider the marketing database a key asset of the organization and database marketing activities as a strategic advan- tage, there are important disadvantages to outsourcing:

♦ Loss of direct control of database management and functioning ♦ Outsourced techniques are not guaranteed to be proprietary ♦ Possible uncertainty of the long-term stability (change of management

and personnel, mergers, solvency, etc.) of the database management vendor

Considering these advantages and disadvantages, the organization may opt to contract with a database management vendor to house and update their database. This relationship works best when both the client and the vendor are very clear on the parameters of their respective responsibilities. The recommended process to select an external database provider is shown below:

Drozdenko05 2/26/02 6:12 PM Page 88

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

Technology, Organization, and Planning 89

♦ Assess internal needs (required whether or not the database build is outsourced)

♦ Formalize needs into a specification for a proposal ♦ Distribute project specification to qualified vendors ♦ Evaluate proposals versus corporate needs and budget ♦ Make selection

Internal Needs Assessment

In this phase, all constituents must be interviewed to determine how a database might serve them. This interview process will help evaluate who the end users of the database will be. In particular, functional areas of Marketing (Advertising/Circulation in a publishing environment), Finance, Corporate Planning, Customer Service, Fulfillment, List Marketing, and Online Marketing should be interviewed to identify the

♦ Data used in each department as part of the job function and its source

♦ Frequency of data update required to perform their job

Once interviews have been conducted, an assessment of the various data sources should be made to determine which data sources used in the organ- ization are specific to customer activity and belong in a customer-level database.

The first step is to identify the needs of all areas. Within the needs assess- ment, it is vital to know what data are used and how they are used. The usage of the information is particularly important, because time-saving efficiencies may be realized as a byproduct of the compilation of data and should be recognized in calculating the database’s ROI (return on investment).

Following this data assessment, the size and number of data feeds can be ascertained. Companies with large IT departments may decide they have the ability to bring the requisite data feeds and information together. Other companies may decide that the information feeds are too varied and they lack the staff to manage the process of building a database on a routine basis.

All constituents should understand that a marketing database is not designed to fully meet the data needs of all departments but rather to be a repository that provides the best view of customer behavior and attribute information. For instance, if a finance function is to assess returns quarterly and this information is provided by reports from the product warehouse via fulfillment, a database can provide this information as well. Both database and warehouse reports should reflect the same level of returns monthly. But a database will provide additional information so that a return profile can be developed for a customer. Conversely, warehouse reports will show total

Drozdenko05 2/26/02 6:12 PM Page 89

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

90 OPTIMAL DATABASE MARKETING

returns and the proportion returned to stock, important from a finance perspective, which may not be captured on a database.

Formalize Needs Into a Specification for a Proposal. To gather external vendor proposals for a specific business needs, a detailed specification must be written and distributed to external vendors. A detailed document designed to aid in the development of proposals is called a request for proposal (RFP). A detailed assessment of database needs is essential or the specification will be out of sync with the actual database requirements.

In addition to identifying all the data sources accessed by information users, it is important to communicate

♦ The size in terms of number of records ♦ Types of records from each data feed (e.g., fulfillment, mail files,

Internet, list enhancement, internal surveys) ♦ How data from various feeds relate to each other ♦ Specific tables that will require maintenance (e.g., product tables,

offer tables, list tables) ♦ Timing of data flow ♦ Changes in business (as it relates to information feeds, e.g., increased

circulation, cancellation cycles) ♦ Requirements for file maintenance such as NCOA

In a specification document, it is important to outline in detail exactly what data will be processed, how they should be processed, and with what frequency they will be processed. In addition, you should specify the analytic needs associated with your database.

Many database vendors have proprietary analytic packages that are used with their database structures. Other database designs are generic enough so that off-the-shelf analytic packages can be used with the database architecture. It is important to include in the specification document the type of analyses required with the database. The proposal from the vendor should therefore address the analytic needs.

Distribute Project Specifications to Qualified Vendors. Following a com- prehensive development of specifications, the RFP document is distributed to vendors. The selection of vendors to provide proposals should be dictated by the following guidelines:

♦ Familiarity with the company based on reputation or industry word of mouth

♦ Vendor’s experience with industry (catalog, Internet, publisher, retail, b-to-b, etc.)

♦ Particular strength in an area of importance to your project ♦ Breadth of services available from a larger organization

Drozdenko05 2/26/02 6:12 PM Page 90

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

For example, if a company marketing upscale products is seeking a database vendor, it makes sense to include a database provider who is affiliated with a list company that specializes in the affluent market. The client who markets upscale products will receive synergies not available from a provider who lacks resources in marketing to the affluent market.

Evaluate Proposals Versus Corporate Needs and Budget. When all pro- posals are returned, they should be compared to each other and the original specification document for completeness and soundness in project planning. Points of comparison include the following:

♦ Timing ♦ Cost ♦ Technology ♦ Software ♦ Service levels ♦ Support team expertise (years, industry, related projects, etc.) ♦ Client training programs ♦ Performance guarantees ♦ Response time to requests ♦ References by existing clients

Attributes of all bidders should be compared side by side and discussed by the management committee charged with making the decision. The vendor proposals should contain all the information requested in the specification. The organization of each proposal will be unique, however, so it may be time consuming and confusing for a nontechnical person to read. A consultant can be a worthwhile resource in providing comparisons across the various dimensions among all vendors.

Make a Selection. In general, the proposals are evaluated and finalists are selected (i.e., some group of proposals are eliminated and others represent- ing a better fit are retained for further consideration). Vendor companies whose proposals are retained may be invited to formally present their results. In this case, a dialogue can occur to identify any questions relative to the proposals. Following this stage, a company should have enough information to select a vendor.

Technology, Organization, and Planning 91

_________________________ Phases of Database Development

Whether built in-house or outside the organization, database development can be complex and time consuming. To expedite development and increase the probability of successful completion, a systematic process has been

Drozdenko05 2/26/02 6:12 PM Page 91

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

proposed. Jackson and Wang (1997, pp. 146–148) recommend the following four phases of database development:

1. System Design: User and data requirements, systems, implementa- tion, costs

2. System development: Physical development and testing

3. System implementation: Live testing in final form

4. Ongoing management

In the initial stage, system requirements are gathered from Marketing and other departments. At that point, IS can propose a system that will meet the requirements and estimate implementation schedules and costs. Marketing must make sure that the requirements they submit are clear and complete.

Once the proposal is approved, the needed hardware and software are configured and tested. Smaller working models or prototypes may be used to test the system. Depending on the organization, Phase 2 may require major acquisitions of hardware and software or only minor modifications to existing systems.

Transition from prototypes to full working systems is the next phase. For new database systems, this transition phase provides an opportunity to evalu- ate the system with real data and usage situations. This “live” testing of the system is even more critical for organizations that currently have large database systems in place. A disruption to database processing during the transition phase could result in a substantial loss in revenues. Therefore, organizations will continue to run the old system along with the new system until all aspects of the new system are checked. During this phase, Marketing should systematically check as many aspects of the system as possible that have an impact on their functioning. Analysis, reporting, access, data updating, coding, tracking, and so on, should be evaluated and reported to IS.

The final phase, ongoing management, is never completed. As we dis- cussed, many aspects of the database need to be updated on a regular basis. The schedule of maintenance should correspond to the system requirements in Phase 1. As the database system evolves over time or business conditions change, new updating requirements may be needed. For example, to meet customer needs in an increasingly competitive environment, more frequent data updating may be required than originally specified.

92 OPTIMAL DATABASE MARKETING

Comments on Technological Development of the Database ____________________________

The process of database development can be expedited if there is cooperation among the functional areas of the business. People from other functional

Drozdenko05 2/26/02 6:12 PM Page 92

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

areas often remark that marketing personnel are not sympathetic to the challenges and limitations of the other functional areas. Marketers could make similar remarks about the other functional areas. Although conflicts between functional areas are often unavoidable, the degree of disruption can be mitigated. Marketers, by clearly defining their database requirements and being receptive to IS’s perspective, should be able to help expedite the development process.

Technology, Organization, and Planning 93

_______________________________________ Chapter Summary Although marketers usually don’t have primary responsibility for the tech- nical development of a database system, they should be aware of basic technology concepts. A range of hardware and software options allows database systems to be adapted to an organization’s requirements. Within these options are certain limitations and advantages that affect how the database is used. Marketers usually want database systems that allow maximum flexibility to add fields as new marketing programs are initiated and allow easy extraction of selected data for timely analysis. The organi- zational considerations in database systems development can be complex and often require a multidisciplinary team to optimize the process. In many organizations, Marketing’s role in database development focuses on detailing their requirements for the system. Therefore, Marketing should strive to provide a clear definition of not only present needs but also anticipated future needs. In addition to detailing the specification for the database system, the organization has to determine who will be responsible for database development. Outsourcing some or all of the development is an option for many organizations. Regardless of the database specifications and development responsibilities, database systems should be implemented in systematic stages in order to minimize disruptions to the organization and allow sufficient testing of elements of the system.

_______________________________________ Review Questions 1. What is Marketing’s primary role in database development?

Provide an example of an organization where Marketing’s role in database development might be more extensive.

2. What are some hardware options for a database system?

3. What are the advantages and disadvantages of using commercial database software packages versus developing database programs from scratch?

4. How is a structured database different from a relational database? What are the advantages and disadvantages of each?

Drozdenko05 2/26/02 6:12 PM Page 93

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed

5. What are some of the organizational and cross-functional factors in database development? Consider outsourcing options.

6. Discuss the implications of phasing in a new or upgraded database system.

94 OPTIMAL DATABASE MARKETING

Drozdenko05 2/26/02 6:12 PM Page 94

C op

yr ig

ht ©

2 00

2. S

A G

E P

ub lic

at io

ns . A

ll rig

ht s

re se

rv ed