Chapter 3

The Forest for the Trees: Where Is the Customer in All That Data?

In This Chapter

Understanding the basic structure of a marketing database

Using independent service providers to clean up your data

Grouping customers into households

Complying with privacy laws and standards

Your success as a database marketer is closely tied to the quality of your marketing database. Marketing databases are strange creatures in the world of information technology. For one thing, they tend to evolve significantly over time. For another, they require a good deal of care and feeding. This means that database marketing is a team effort and requires support from a number of different departments in your company.

In this chapter, I lay out some basic concepts and terminology that will help you understand your database environment. This information will also help you communicate effectively with your partners across your organization. I also outline your key responsibilities in holding up your end of the database bargain.

A Marketing Database Is About Your Customers

Your ultimate goal as a database marketer is to increase revenues through direct communication with your customers. The more you know about your customers, the more effectively you can communicate.

No matter what business you’re in, you’re surrounded by data. Sales data, website data, financial data, product data — the list goes on. The systems that contain this data were designed to keep your business running. Cash registers, accounting systems, online booking engines, and movie ticket kiosks are all examples of such systems. You will hear them referred to as transactional or operational systems. The idea is that they are meant to efficiently manage the day-to-day operation of your business.

What makes a marketing database ­different

What your operational systems are not designed to do is answer the kind of questions that you, as a database marketer, are asking. Your accounting department, for example, could quickly tell you how many widgets your company sold between 3 and 5 o’clock on Monday. What they probably can’t do is tell you much about who bought them. That’s your job, in a nutshell.

The questions that naturally occur to you are not top of mind to the folks on the operational front lines. But even if they were, the data may simply not be available to them. Operational systems tend to operate in data silos. This means that they only store the data that is directly relevant to their specific function.

A movie ticket kiosk doesn’t need to know how far a moviegoer lives from the theater. But if you were trying to decide how widely to mail a movie ticket offer, you would certainly care. You might also care how old they are and how many kids they have. These things could affect how you might communicate with them.

What you need is a database that’s organized around the customer. And that means pulling together data from all (or maybe just some) of these data silos and organizing it in a way that serves your marketing needs. This is the basic notion behind a marketing database.

In your discussions with your IT partners, you will almost certainly hear someone refer to the ETL process (see Figure 3-1). ETL stands for extract, transform, load and is in reality three distinct sets of processes:

Extract: Pulling data out of the operational data silos. The most important thing IT will want to know from you is how often extraction should be done. For example, do you need the data every night? Or is weekly enough?

Transform: Anything that is done to clean up, standardize, consolidate, or otherwise make the data ready for prime time. Whenever a change is made to the database, you should be asked to approve what is done here.

Load: The actual loading of the database.

Illustration by Wiley, Composition Services Graphics

Figure 3-1: A typical depiction of the ETL process.

What’s in a marketing database

The specific data that best suits your needs is highly dependent on both your business and your systems environment. But there are some categories of data — referred to usually as subject areas — that are fairly universal in marketing databases. The following list gives a brief description of the most common subject areas:

Address data: Address means more than just postal address. This category includes e-mail addresses, telephone numbers, cellphone numbers, and any other contact information that you might use. This category also includes customer contact preferences and records customer requests to opt out of communications.

Transaction data: Includes what you know about what the customer has bought from you and when. It includes information about where and how each purchase was made — for example, online versus at a store. Because there tends to be a lot of it, transaction data almost always needs to be summarized to make it manageable. I address this topic in more detail in Chapter 8.

Demographics: Age, income, and marital status are typical examples of demographic data. A staggering array of data is available from third-party vendors. These vendors are typically quite willing to work with you to identify the data that is most relevant to your business. I discuss the use of demographic data in detail in Chapter 7 and I also point out some companies that sell demographic data in Chapter 19.

Internet/mobile device data: Virtual profiles are a valuable source of customer information. Because customers frequently create these profiles themselves, the profiles contain information that comes directly from the customer. This self-reported data can be some of the best quality data in your database.

Shopper data: This category is similar to the transaction category. The difference is that these records didn’t end in purchases. Browsing your website, requesting a catalog, and signing up for a newsletter are all considered shopping behaviors.

Campaign history: An absolutely critical part of your database, this is a record of your marketing campaigns and who responded to them.

One thing to keep in mind is that data is being updated all the time. But not all data is updated on the same schedule. It’s important to be aware of the “freshness” of your data. As I discuss later in this chapter, addresses can become outdated. In Chapter 16, I point out the importance of knowing how fresh your data is when you analyze marketing campaign results.

Your database needs to keep track of when data has been updated. This is standard operating procedure in database management. Virtually all database software provides the ability to “stamp” every piece of data with the date and time that it enters the database. These date and time stamps provide you with the ability to determine not just what’s going on with your customers, but when. As I discuss in Chapter 10, timing your communications based on customer behavior is a powerful marketing technique.

How a marketing database is organized

Connecting all this data to a customer is not always easy. Sometimes it’s impossible. Often, customer information isn’t stored in your operational systems. If you work for a bank, this is not a big problem. At banks all transactions are tied to accounts, which are tied to customer records. The IRS, among others, cares very deeply that this is so. But if you work in retail — say, a grocery store — you don’t have this luxury. I can walk in and pay cash for a gallon of milk, and you’ll have no idea I was there.

Your ability to tie your company’s data to individual customers is highly dependent on the nature of your business and the details of your information systems.

There are, however, some tricks of the trade that can help you make some of those connections. In the case of grocery stores, it has become commonplace to get customers to sign up for rewards cards. The grocery store I use gives me discounts on gasoline for every 50 dollars I spend. This makes me quite willing to let them scan my rewards card every time I shop there. Because they mailed me my rewards card when I signed up, the rewards card has the effect of tying all my purchases back to me. In Chapter 18, I divulge an assortment of tricks to help connect your data to your customers.

A marketing database is not typically limited to just data from within your company. Third-party marketing service providers can supply you with extremely useful data about your customers. I address the nature of this third-party data in greater detail in Chapter 7.

These providers are able to give you significant insight because they collect data from a wide array of sources across many industries. Definitely familiarize yourself with the major data providers and keep yourself up to date on new offerings. In Chapter 19, I point you to some of the big players in the demographic data arena.

The key to an effective marketing database is that everything revolves around the customer. The litmus test for whether to include some piece of data is this: “Can I connect the data to a customer?” You want to make the customer the star of the show. A very common and intuitive way to conceptualize this idea is illustrated in Figure 3-2. This figure, known as a star schema, shows the customer at the center of all the data. Data from various sources is connected to the customer like spokes in a wheel.

Illustration by Wiley, Composition Services Graphics

Figure 3-2: A simple star schema organization of customer data puts the customer at the center.

Some Assembly Required: Building the Customer Record

So the customer is the star of the show. But it’s not always easy to get the customer to center stage. At the most basic level, a customer is usually defined by a name and an address. Creating a customer record means matching names and addresses from different systems. You want to match the Elizabeth Smith of 123 Main St. who registered online to the Elizabeth Smith of 123 Main St. who signed up to receive your quarterly newsletter.

Sounds simple enough, right? The problem is that the data is rarely that clean.

Cleaning up addresses

Elizabeth may have registered as Elizabeth Smith online two years ago and mistyped her address as 123 Mane St. She may have more recently requested the newsletter be delivered to Elizabeth Smith at 123 Main Street.

It’s no problem for you to recognize that St. and Street are a match. It’s also fairly obvious that Mane is a typo and means Main. But you don’t have the time to sort through all your address data. You need a computer to recognize this for you. And explaining the obvious to a computer is anything but simple.

Luckily this problem has been solved for you. Commercially available software packages will standardize your name and address data for you. They actually use the official U.S. Postal Service address database to identify invalid addresses. You can even purchase services that will perform these functions as data is being entered into your systems. For example, when a customer registers on your website, the software will not let them enter an invalid address.

The USPS website at www.usps.com lists software packages that it has evaluated and approved. Choosing the software that's best suited to your needs is more an IT function than a marketing function. The software is system specific. It depends on your database and operating system environment. But the good news is that you may be able to get your IT department to pick up the tab in its budget because software is essential to managing the marketing database.

Address standardization and validation are essential to building and maintaining a useful customer database. Software packages that perform these functions come with various bells and whistles for various prices. You may not need the high-performance sedan of software. But you can’t afford not to have the basic model.

Updating addresses

Even after your name and address data have been standardized, you still have a potential problem: Customers move. The customer’s address may not be current in all of your company’s systems. In other words, you may have several different addresses that are associated with the same customer. If you’re not careful, these different addresses can create the impression that you have several different customers.

I set up an online banking account a few years ago. I had just moved to Florida and had never done business with Bank XYZ — or so I thought. At some point, I had to enter my social security number, and the website immediately pre-filled my name and address data. I immediately realized it was the wrong address, but it looked vaguely familiar. Finally it dawned on me that I had lived at that address 15 years earlier while I was in college. Bank XYZ had bought the bank I used in college and were using the old bank’s online banking system. That system had not heard from me since I left college.

Fortunately this problem is also easy to solve. The Postal Service maintains a database of address changes. This National Change of Address database is more commonly known by its acronym NCOA. Many of the companies that offer address standardization services also perform address updates using this database. Using an NCOA service is standard practice in database marketing.

Once you have your addresses standardized, validated, and current, you’re in a position to create a central customer record. You have identified the star and are ready to get on with the show.

Marketing Is a Family Affair: Understanding the Household

Understanding a customer’s household is a critical component of your database marketing efforts. Grouping customers together into “family units” is known as householding. In this section, I explain in general terms how this is done. More importantly, I examine some reasons why it is done.

What is a household?

In its most rudimentary (and naive) form, a household is a group of people with the same last name living at the same address. This definition is based on an outdated 1950s Leave It To Beaver view of the nuclear family. This view is clearly problematic in the real world. Many couples, both wed and unwed, do not share the same last name, for example.

What you’re concerned with as a marketer is identifying domestic arrangements that involve some sort of mutual financial or budgetary decision making. Two college students sharing a dorm room do not have the same budgetary connection as a young married couple planning for their first child. Thus you would not generally be inclined to consider a dorm room a household. The young married couple’s apartment, however, would be.

The nuts and bolts of householding are definitely not simple. If you don’t use an outside service provider for anything else, hire one to do your householding for you. I list some of the larger providers of these services in Chapter 19. These service providers have access to data on virtually every household in the country from a wide variety of sources. They can see connections that you would never be able to spot in your own data.

Why the household is so important

One of the primary reasons for householding is to keep mail costs down. There is typically no reason to send separate marketing offers to every single person in a household. Especially when the offer involves something the household is only going to need once. For example, a family is typically only going to buy one local newspaper subscription. It would be senseless to keep offering everyone in the household the same subscription after one of them had bought it.

Generally, one communication per household is a good rule of thumb for your marketing campaigns. This rule is most important for more expensive campaigns, like first class direct mail. It’s less critical for very inexpensive campaigns, such as those delivered by e-mail.

Because you’re only mailing once to each household, you have a decision to make: Who gets the letter? You can sidestep this issue by simply mailing to the Smith Family. There are certainly cases where this makes sense. But it is also the case that such communications are more frequently discarded. Many marketers decide to flag one member of each household as the designated head of household. This is the person to whom marketing communications are addressed.

Contrary to common preconceptions, it has been fairly clear to marketers for a long time that more often than not, it is the wife, not the husband, who drives purchase decisions. For this reason, designating the eldest female as the head of household has become fairly common. Depending on your business, though, you may decide it makes more sense to communicate with a different member of the household.

Mailing just once to a household also reduces the risk of double counting responses. Say you mail multiple pieces to the same household, and the household actually makes a purchase. A common mistake is to count that purchase separately against each piece of mail that you sent to the household. In that case, you dramatically overstate the success of your campaign. I examine campaign evaluation in detail in Chapter 15.

The household is important for yet another reason. You are interested in understanding your customers as well as you can. A great deal of information that is available to you from outside vendors is actually household-level information. The famous Nielson TV ratings are gathered by household. Census data is gathered at the household level. Income data is generally reported at the household level.

You may decide to purchase household-level data about your customers. If so, it’s important that you purchase data that’s relevant to your business and your situation. The accuracy of each individual piece of purchased data varies greatly according to where it comes from. It’s a good idea to compile the best data from multiple providers based on their ability to provide quality and accuracy for your unique customer base. You can do that by providing a sample of the database to several vendors and comparing the data that they provide.

Growing Your Customer Base: Prospective Customers

Your marketing database is about your customers — not just your current customers, but also your potential customers, otherwise known as prospects. It’s not enough to keep your current customers happy and doing business with you. You need new customers to grow. You’re always looking for prospective customers. Knowing who has shown some interest in your products gives you a head start in this search.

Online shoppers

Your website is a natural place to look. People do a lot of shopping online. By encouraging them to register while they’re exploring your website, you can create an ongoing stream of prospective customers.

It’s sometimes surprising how little incentive many people need to register on your website. You don’t necessarily need to offer them special discounts. Sometimes it’s as simple as offering them access to a planning tool or other convenience in exchange for the registration information. The idea is that you’re trading information with your potential customer. The power of this approach is that you get more than the customer’s contact information. You now know that the customer is interested in you.

My wife and I recently decided to replace our living room furniture. Because this represented a significant investment, we obviously wanted to put some thought into it. After shopping around on a few websites, we stumbled onto one that offered a really slick little application that allowed us to visualize the company’s furniture in our living room.

It required us to register on their website to use it. But this registration actually served a purpose for us. It allowed us to store the dimensions of our living room on their site so that we could come back to it again and again. It also allowed us to store prospective furniture selections and layouts for comparison.

The company ultimately contacted us and set up a showroom appointment. We were already so pleased with their service that we were practically sold before we walked in. (If you're interested, check out the My Projects tab at www.ethanallan.com.)

Automobile companies employ a similar technique on their websites. They allow customers to select features and packages. Then, in exchange for their registration, they display a full color photo of the resulting car, complete with price and availability.

Many hotels and travel companies offer vacation-planning tools in exchange for registration. Maps, videos of the destination, and suggested itineraries are all examples of companies trading information with their potential customers.

In trying to generate website registrations, being reasonable about the amount of data you are trying to collect is important. Asking prospects to fill out an extensive form with their name, address, telephone number, and other personal information can be off-putting. At the very least, it takes away from whatever convenience you’ve told them could be gained by registering. Keep it simple. Sometimes just requesting the customer’s e-mail address can be enough to get started. It’s very easy to collect more detailed information when they decide to make a purchase. At that point, they need to provide that information to pay and receive shipment of their purchase anyway.

Call center data

If your company operates a call center, it can be another valuable source of information on customers’ interests. Call centers are used for a variety of reasons. Customers call in to place orders, submit service requests, make ­payments, and complain about problems. Regardless of the reason, the call center agents are in a unique position to gather information about callers’ interests and situations. They are actually talking directly with the customers.

Though it takes an investment in technology, much of this information can be captured and fed into your marketing database. It’s become standard operating procedure for medium to large companies to implement call center systems designed to capture customer data. These systems also feed data to the call center agents from other sources, including marketing databases.

This type of call center system is often called customer relationship management (CRM) or sometimes sales force automation. Oracle/Siebel and SAP are two of the biggest players in this market. But there are literally hundreds of other companies that have products in the market across a broad spectrum of prices. Once again, the Direct Marketing Association's website, www.thedma.org, is a good place to start.

Some other sources of prospects

Newsletters are another useful tool in generating prospective customer lists. The principle is the same here as with web registrations. You give the customer something of value to them — namely, news about upcoming events and offers. The customer gives you their contact information along with an implicit acknowledgment that they have an interest in your product.

You may also have some opportunities to identify prospective customers in your internal source systems. The nature of these opportunities is highly dependent on your particular business and on your company’s system infrastructure. But you should certainly explore these opportunities. At the very least, have some informal conversations with some of the folks who interact directly with your customers. I give some examples of how to get more useful data from your internal systems in Chapter 16.

When all else fails, you have another option: Buy a list. There are literally hundreds, perhaps thousands, of companies that are willing to sell you lists of prospective customers. Many of them will work with you to define a prospective customer profile and provide only names that meet this profile. In The Part of Tens, I point out the importance of belonging to the Direct Marketing Association (www.thedma.org). The website is a good place to find list providers.

The quality of purchased lists varies dramatically among list providers. It’s important to understand where a list provider gets its data and how detailed it is. If you get vague answers to your questions on this subject, it’s a red flag. Many providers brag about having very detailed age and income data, for example. But in some cases, this data is really only reflective of averages by neighborhood. All households in the neighborhood carry exactly the same description. I address this topic in more detail in Chapters 7 and 19.

One effective way to evaluate a list provider’s data is to hand them a sample of names from your own database. Ask them to pick out those names that fit the customer profile you want to purchase. By comparing their results to the information you have in your database, you can get a good sense of how well they are doing.

Regardless of where you find them, converting prospects into customers is a core part of your job as a database marketer. I’ve found this aspect of the job to be quite fun. So much useful data can be found outside your company’s operational systems. Going that route gives you the freedom to get creative about ways to engage your potential customers without needing to worry too much about impacting the core business systems. And given the pace of change in our world of electronic media, you will be continually coming up with new ideas.

No Trespassing! Respecting Your Customers’ Privacy

Everyone is concerned to some extent about keeping their personal data personal. Identity theft is a potential nightmare lurking out there for all of us. You wouldn’t think of storing your credit cards, social security card, birth certificate, or passport in the glove box of an unlocked car. If you’re like me, most of those items are safely locked away.

Beyond your concerns about being fleeced by an identity thief, there are many things you simply don’t care to share with the world. You don’t publish your bank or investment records. You don’t post your medical records on social media sites. Simply put, you value your privacy.

Your customers feel the same way that you do about their personal data. It’s your responsibility to be your customer’s advocate with respect to these concerns. You should treat your customer’s personal data with the same respect that you treat your own.

Protecting customer data

Your customer data is a valuable corporate asset. You need to make sure that security measures are in place to protect that asset. Doing so is to some extent a technical exercise. But hacking into your computers system isn’t the only, or even the most common, way for security to be breached.

Using technology to protect customer data

In part, data security means ensuring that technical measures are in place to prevent your customer data from unauthorized access. Your IT department will help a great deal in this effort. Network security, database encryption, and robust password standards are all necessary tools in the effort to safeguard data.

For starters, you need to store your data on secure servers that can only be accessed by authorized users. That access needs to be restricted using the full breadth of your network security system. Passwords need to meet security standards. An employee shouldn’t be able to use their dog’s name as a password, for example.

Your database itself needs to be password-protected as well. Virtually all database management systems offer the ability to do this. They also offer the ability to restrict access to individual variables in the database. This extra level of restriction should be considered for particularly sensitive information like credit-card numbers or social security numbers as well as names and addresses.

Beyond restricting access to sensitive personal data about individual customers, you should encrypt this data. Even if the data is somehow downloaded, encryption gives you another level of protection against the data being misused. Encrypting this sensitive information is not generally a problem for your analysis efforts. Beyond zip codes, you aren’t going to be using addresses and credit-card numbers in your analysis and reporting efforts.

You also need to encrypt any data that you share with your marketing service providers. You’ll need to send lists to your mail service providers, for example. Whether you’re overnighting them on a CD-ROM or sending them electronically, they need to be encrypted. Vendors will typically have secure file transfer infrastructure already set up.

It’s generally not a good idea to send sensitive customer data, encrypted or not, via e-mail. It’s far too easy for e-mails to be intercepted or inadvertently forwarded. Using e-mail limits your security to a single layer of encryption, which can be more easily broken than the multiple layers provided by your servers and your secure file transfer channel on top of the file encryption.

---

---

Sharing customer data

There are a variety of reasons why you may consider sharing customer data with third parties. When deciding what data to share with whom, you need to be aware that consumers are getting more and more sensitive about having their personal data used for marketing purposes. And as I point out in this section, you need to communicate clearly to your customers what you’re doing with their personal data. There are several different scenarios related to data sharing.

For starters, you need to share your data with vendors that provide marketing services. The companies that provide address cleansing services, householding, demographic appends, and the mail houses that execute your campaigns all need access to your data. This is just standard operating procedure in database marketing.

In large companies — particularly companies that have a variety of different business units — customer data is often shared among these different subsidiaries. Because all these business units are technically part of the same company, this type of data sharing is not particularly problematic. But be aware that customers don’t always recognize all of your brands as being part of the same company. For example, many consumers don’t realize that the Italian restaurant Carrabba’s and Outback Steakhouse are owned by the same company.

Companies may also decide to share data with corporate partners. An extremely common example is when a company partners with a bank to offer a cobranded credit card. Many of the airlines have such programs. Here data sharing is almost a necessity if the corporate relationship is to be successful.

Finally, companies sometimes choose to sell customer data. I’m not talking here about marketing service providers who aggregate data and sell it to marketers. They’re not providing data on their customers. I’m talking about selling your customer lists to third parties. This is the sort of data sharing that consumers (and legislatures) are most sensitive about.

The decision to sell your customer data to third parties depends a great deal on your business model. In my experience, most companies decide not to do this. But a growing number of companies, particularly in the digital media arena, have a business model partly based on providing marketing services based on their customer data. If you do decide to sell customer data, you need to stay up to date on legislation related to consumer privacy, as discussed later in this chapter.

Keeping your customer informed

Respecting your customer’s privacy is about more than just data security. It’s also about transparency. You need to communicate with your customer about how you collect and secure their data. The customer also needs to understand how you’re using their data. Many companies choose to sell or share their data with corporate partners, as mentioned in the preceding section. Whether and how you do this needs to be transparent to the customer as well.

All this information needs to be well documented. What you need is a privacy policy. This policy should be developed in conjunction with your legal department and senior management. It should be clearly communicated within your organization to ensure that it is followed. It should also be made easily available to your customers. At the very least, your website should contain a Privacy Policy link that allows users to review the complete policy.

Privacy policies are standard practice in virtually all medium to large companies. A simple web search can provide you with a host of examples that are specific to your industry. You probably get a privacy policy update letter from some company or other a couple of times a month.

Some legal considerations

Privacy policies are often not optional, but legally required. There are several industries whose data and data usage are governed by federal legislation. And new legislation continues to be proposed. It is critical that you understand your industry’s legal environment.

Differences by industry

One industry that has come under broad scrutiny by the government is the financial services industry. The Gramm-Leach-Bliley Act puts significant restrictions on the sharing and use of customer data by banks and other companies involved in consumer finance.

Another industry that is highly regulated is the health care industry. The Health Insurance Portability and Accountability Act is a federal law governing data related to a person’s health care. Data on children is highly sensitive as well. There are laws on the books and in the pipeline governing the gathering, storage, and sharing of information about young people.

In addition to federal legislation, some state governments have passed laws regarding the use of consumer data. Also, if your company does business outside the United States you’ll need to understand the legal environments there as well. The United States is actually one of the least restrictive — anybody thinking about using their database for international marketing needs first to consult counsel and understand the regulations.

Differences by channel

In addition to being industry specific, both existing and proposed legislation regarding privacy differs by marketing channel. As I discuss in Chapter 4, you need to give your customers the opportunity to request that you not market to them. Your legal obligations regarding these opt-outs differ according to how you’re communicating. Different legislation applies to telemarketing than to e-mail, for example.

But your obligations can also vary according to how data is collected. The widespread use of mobile devices has gotten the attention of federal and state legislatures throughout the U.S. The privacy issues related to mobile devices stem primarily from the real-time location data available from those devices.

One concern revolves around transparency. Many people simply don't realize what information they're broadcasting from their mobile devices. And because location is an essential part of many mobile apps, these broadcasts are often automatically turned on. Another concern relates to the privacy and safety of children. The Children's Online Privacy Protection Act (COPPA) was first passed in the late 1990s. This law aims to limit the data that websites collect from children under 13 years old. This law is actively enforced by the FTC and has resulted in a long list of companies that have been fined for breaching it. (For more information on COPPA, visit the FTC website at www.business.ftc.gov/privacy-and-security/childrens-privacy.)

More legislation is brewing around mobile device data. More and more children are carrying smartphones everywhere they go. As I write this, legislation is being considered in Washington to address concerns specifically related to mobile devices and location data. These concerns are legitimate. And the need to protect children is one of the few things Americans all agree on. I will be shocked if some sort of legislation isn’t passed. Stay tuned.

Develop a relationship with the lawyer in your company who is responsible for privacy compliance. But also stay abreast of what is happening in your industry with regard to privacy legislation and especially how other companies are responding to it. You want to comply with relevant legislation, but you also don’t want to unduly limit your ability to use your customer data.

Privacy compliance is not the most exciting aspect of your job as a database marketer. And it’s certainly not the most convenient. But it is necessary. As time goes on, signs point to it becoming a larger, not smaller, part of your responsibilities. Privacy is also not the only aspect of your job that’s guided by legislation. There is also legislation governing your need to allow customers to opt out of hearing from you. Chapter 4 addresses this in more detail.