How a Relational
Database Helps Marketers
To do database marketing properly, you want to collect data on your customers, and use that data to measure their lifetime value and to communicate with them. The messages can be personalized because you have stored information properly so that the database can enable you to have meaningful dialog with hundreds of thousands of customers, just as the old time corner grocers had with their individual customers.
Database marketing today is based on relational databases maintained on servers (or in the case of very small databases, on PCs.). Databases used to be stored on mainframes, but these are not acceptable for modern conditions. To get started designing a relational database, most companies use a Needs Assessment, which is a two day exercise interviewing all concerned: marketers, sales, management, IT, and vendors. The goal is to determine what the marketing program will be for the next year, and the data available and necessary to support that program. After the two days, the database design team spends the next 28 days creating the database structure.
A relational database consists of fields, tables, and records. A field is the smallest structure (the atom) of the relational database. Examples are: A last name, a date, a dollar amount, a product number. There may be a hundred fields in a database. Properly designed fields have only one value: Wrong: Arthur Hughes. Right: two fields 1) First Name 2) Last Name. They should not contain data that changes or calculated values: Wrong: Age. Right: Date of Birth.
Fields are grouped into Tables. A Table is the chief structure in a relational database. A table always represents a single specific subject which can be either an object or an event. Object example: Customers. Event Example: Orders.
Keys are special fields. Every table contains a primary key field and may have one or more foreign key fields. A primary key field for the Customers table would be the Customer Number. A foreign key field for the same table would be an Order Number field (which is the primary key in the Orders table). The other fields in the Customer Table would be data fields, such as name, street, city, state, zip, etc.
Records are the structures in a table that represent unique instances of the subject of the table. An example is Arthur Hughes, one of the records in the Customers table. A purchase that Arthur Hughes made on August 12th could be an example of a record in the Orders Table.
Relationships are connections established between pairs of tables.
This diagram shows the relationships between three tables. The relationship is between their Primary Keys (PK) and Foreign Keys (FK).
Why should you care?
Why is all this technical detail important to database marketers? Because, in most cases today, their databases are not designed properly. They are old fashioned mainframe flat files. The update process is very slow. Adding or changing data in a flat file is a difficult process. What are the advantages of a relational format to a marketer? A modern relational database:
Because marketers do not know how relational databases are constructed, they tend to be dominated by IT professionals who tell them what can and cannot be done. IT professionals explain how difficult it is for them to do whatever the marketers want done. To be an effective database marketer today, therefore, marketers must learn some of the basics described above so they can understand what IT is saying, and get out from under their thumbs. They must look at the tables set up by their IT design team to be sure that they are well constructed to carry out their marketing programs.
The Business Rules are central to marketing use of a relational database. They are built into a relational database structure. Business rules can dictate not only the correctness of data stored in records (dates in correct format, product IDís are correct, zip codes consistent with City and State, for example), but they can also permit marketers to create event driven communications. What are these?
A view is really a virtual table. It usually contains data from many tables arranged like a report or a screen shot. Typically, you may want your customer service people to see a view of each customer while they are talking to them on the phone. The view includes not only the customerís name and address, but their most recent purchases, their total purchases, and, probably, their complaints or preferences. For you, as a marketer, the view is probably the most useful part of the database. The view can contain calculated values. Date of Birth is converted into age in a typical view. Total number of orders and dollars purchased in the past twelve months is another useful field to show in a view. Views make the database come alive for you, for customer service, and for management.
Access to your customer information
Once you have a relational database set up, how do you use it for marketing? In most modern applications, you use access software like Brio Query or TopDog (there are many others) that enable you to work with your database over the web. You may have a dozen marketers each of which needs only web access and a (free) browser to look simultaneously at your customer database and do:
When you log on to this software, you will see on your screen the tables in your relational database. They look like the example shown above. You can pick out any fields of interest to you and create queries that result in reports or customer segments. The results should be ready in a few minutes. If you donít like what you see, you can redo it. You will have the power to get information out of your customer database, and create action programs that communicate with your customers and get them to buy more, visit your retailers more often, and be more loyal.
Where can you learn more? There is an excellent book by Michael Hernandez called Database Design for Mere Mortals (Addison Wesley 1997) which is easy to read. It will make you an expert.