Chapter 13
Using Customer Data Online
In This Chapter
Understanding how to use e-mail effectively
Serving up web content based on customer information
Recognizing your customer online
Being visible through search engines
The so-called virtual world has evolved at lightning speed over the last couple of decades. Search engines are often the first place consumers go when they want to shop for some product. Google has become a verb in virtually every language. E-commerce has overtaken traditional in-store purchases in many industries. I haven’t been into a bank branch in years.
The virtual business world is vastly different than the brick-and-mortar world. Change happens much faster — you can set up a website in a matter of hours. Website content is constantly undated, and functionality is enhanced all the time. Perhaps the most significant difference, though, is that websites serve as marketing and advertising platforms as well as cash registers that process sales transactions.
The latest big change online is the flood of mobile devices. The Internet has gone mobile big time. This shift creates both challenges and opportunities for marketers. The world of smartphones and tablets is, if possible, even more dynamic than that of the traditional web.
Because of their ever changing nature, the discipline(s) of online and mobile marketing have developed largely outside other traditional marketing disciplines like advertising and direct marketing. New decisions are made on a daily basis, and the overall culture in these fields is one of experimentation.
I knew an Internet guy a few years back who had a sign on his door that read “Informed trial and error outperforms detailed planning every time.” And that rings true. Waiting until you’re sure you’ve got everything perfect is a sure way to fall behind.
Another feature of the online world is the blurring of the distinction between marketing and operations — particularly sales. As a database marketer, you don’t have to pay attention to what is happening at the checkout counter of your stores every day. But your online team does. Typically they direct both marketing and e-commerce because the two things are so tightly intertwined.
This chapter introduces you to some basic ideas related to the online world that touch on customer data. I don’t have room to address online and mobile marketing in great detail here. If you’d like to explore this topic in more detail, check out Web Marketing For Dummies (Wiley, 2012).
There’s More to E-mail Than You Might Think
When e-mail first became a popular marketing channel, it started as an in-expensive alternative to direct mail. A direct marketing campaign using e-mails costs a small fraction of what printing and postage cost. In the early days, e-mail campaigns followed the same general principles of more traditional mail campaigns. They were designed and executed following the same mass mailing model.
The use of e-mail in marketing has become a good deal more sophisticated over the years. Due in part to the negative reaction of users to the early overuse and abuse of e-mail addresses (called spamming), marketers now have to be much more careful about how and when they use e-mail to communicate. At the same time, they’ve developed more and more insightful ways of tracking the success of their e-mail campaigns.
Chapters 10 and 11 talk about some of the ways that marketers use e-mail in their database marketing campaigns. Here I discuss some aspects of e-mail marketing related to collecting and managing your e-mail addresses. I also talk about how to track responses to e-mail campaigns.
Understanding how customers deal with spam
Because mass mailings through e-mail are so inexpensive, these campaigns don’t need to have very high response rates to generate sales. It was this fact that led inevitably to the abuse of the channel. People’s distaste for spam or unsolicited e-mail, especially in large volumes, led to all kinds of technical and legal developments designed to block e-mail solicitations.
In other words, people have gotten to the point where they simply tune out much of the marketing related e-mail that comes to them. They can flag an e-mail as spam and never see another e-mail from that sender again. E-mail service providers block e-mails all the time.
My wife and I have a separate e-mail account that we use for only one thing. We make a lot of purchases online, and they invariably require us to provide an e-mail address. Because merchants need an actual e-mail address (yes, websites can check this), we created what amounts to a dummy account to receive the inevitable follow up e-mails. We reserve our real e-mail address for companies that we frequently do business with. This is becoming a common tactic for online consumers. I talked recently with someone in the online marketing business who suggested that as many as 15 percent of the e-mail addresses in marketing databases are being used this way.
In Chapter 4, I discuss allowing your customers to opt out of hearing from you — this is standard practice when you collect a customer’s e-mail address online. Usually it involves offering a check box for them to indicate whether they want to receive special offers and other marketing communications from you in the future.
Respect their decision. You are actually legally mandated to do so per the CAN-SPAM law mentioned in Chapter 4.
There is a loophole in the CAN-SPAM law which allows you to communicate with existing customers. If the communication is related to your relationship with them, you can use e-mail. Shipping information on a recent purchase is an example of such an acceptable use. Confirming a password change is another.
Collecting e-mail effectively
You will no doubt want to send e-mails as part of your traditional direct-marketing campaigns. They are very effective as follow-up reminders to special offers that you may mail out, for example. An e-mail telling your customers that they only have three days left to take advantage of a discount not only serves as a reminder but creates a sense of urgency.
But e-mail is only effective if it gets read. You want to be careful about how often you use this channel. All it takes is a single mouse click, and a customer can block your e-mails and never see them again. In Chapter 5 I talk about the importance of developing an explicit contact-management policy to avoid overcommunicating with your customers.
All e-mail addresses are not created equal. You want to focus on quality over quantity. This means you want to be sure your customer really wants to hear from you before including them in a marketing campaign.
As I discuss in Chapter 11, offering newsletters, product guides, or other informational material can be an effective way of collecting e-mail (or home) address information from your customers or prospects. When a customer makes the request, they are implicitly opting in to hearing from you. But this interest may only be related to the particular information you are offering.
You can improve the quality of your e-mail lists by taking this opt-in process one step further. You can require a second, more explicit, opt in from the requester. Typically this is done by sending an e-mail to confirm the request before sending out the requested information. This double opt-in, as it’s called, has the additional advantage of confirming that the e-mail address is real and is being checked by the customer.
Over time, your database has accumulated e-mail addresses that have been collected in a wide variety of ways. The quality of those addresses also varies widely. Some may no longer be active. Some are opted in. Some are opted out. You will even find some that have done both. Still others never were given the opportunity.
A one-time mass-mailing version of the double opt-in approach can be used to clean up your e-mail address list. By clean up, I mean dramatically pare down the list. In addition to ensuring compliance with CAN-SPAM, doing this occasionally confers a few other advantages:
It improves response rates.
It lets you purge e-mail addresses that are no longer active.
It keeps you out of trouble with your e-mail service provider.
E-mail service providers are extremely sensitive to being used as vehicles for spammers. They are the ones holding the smoking gun if the Federal Trade Commission should come calling about CAN-SPAM violations. If a large number of your customers are reporting your e-mails as spam, then you will eventually be dropped by your provider.
Speaking of e-mail service providers, choosing one is among the most important decisions you will make. You’ll rely heavily on your ESP — not just to execute e-mail campaigns, but to ensure CAN-SPAM compliance and e-mail address deliverability. The ESP will also provide reporting on which e-mails have been viewed and which customers have clicked links in them. The best ESPs provide this information in real time on the web. A good ESP can also provide this data back to you at the customer level so that you can incorporate it into your database for analytic purposes. Because you rely so heavily on your ESP, you should get proposals, including client recommendations, from several providers and evaluate them carefully before making a decision.
Tracking e-mail effectiveness
As I say in Chapter 2, it’s important for marketing campaigns to have a clear call to action. You need to communicate exactly what you want them to do. In the case of e-mail campaigns, you typically direct the customer to your website. In fact, most campaigns contain a link to your website in the e-mail itself.
The simple way to think about responses to e-mail campaigns is that the customer proceeds in a straightforward fashion. They click the e-mail. They read the e-mail and become interested. They proceed to your website and potentially end up making a purchase.
I discuss the purchase piece of the puzzle later in this chapter. In this section, I want to talk a little about the click part.
Who is getting to your website? Open and click-through rates
Customer behavior is actually a good deal more complex than a nice linear progression from clicking your e-mail to purchasing. They may read your e-mail on their smartphone and wait until they get home to browse your website on a larger computer screen. They may click the link to your site but visit it several more times before making a purchase. Some customers may be registered on your website and others not. Customers can take vastly different paths to finally making a purchase.
When analyzing the success of an e-mail campaign, one thing you will want to know is what percentage of people who received the e-mail actually opened it. This is known as an open rate. Of the users who do open your e-mail, you also want to know what percentage actually clicked the link to your website. This is called the click-through rate.
Your e-mail service provider will be able to tell you both these things. Open rates and click-through rates are standard metrics in e-mail marketing. Even these simple-sounding methods can be a little tricky, though. You — or more accurately, your service provider — need to avoid overcounting. Reporting of these rates must take into account multiple views of the e-mail as well as multiple clicks on the link.
What are they doing on your website? View-through analysis
More recently marketers have gotten interested in diving deeper into what customers are doing after they click a web link. This sort of analysis is known as view-through analysis. Be aware, though, that this term means slightly different things to different people.
If you’re running an online advertising campaign involving streaming video, you’ll obviously care how many clicks you get asking to view the video. But more importantly, you’ll be concerned with what percentage of those viewers actually watched the whole video. This second number is referred to as a view-through rate.
The term view-through is also used in the context of website traffic analysis. In the case of e-mail campaign response analysis, it is an extension of click-through analysis. After a user lands on your website, they have a lot of options. They can search for a specific product, for example. They can browse products by price. They can simply wander around exploring various pages. Or they can leave.
Understanding view-through behavior can be very helpful in designing future e-mail campaigns. In particular, knowing what products your customers are searching for and viewing helps you to focus future messages on what’s most relevant. In Chapter 11, I discuss the importance of understanding barriers to purchase. What is it that’s preventing the customer from buying? View-through analysis can be very helpful here as well. When a customer abandons your website, you can tell what they last looked at. You can also tell where they went.
Many websites, particularly ones that sell retail products, allow you to view products by price ranges. A user might be looking at refrigerators in the lowest price category and then immediately jumps to a competitor’s website. This would be a good indication that you have a price-sensitive consumer. You may decide to offer this customer a discount on a higher-quality model. The assumption is that the customer didn’t see anything in their price range that was satisfactory. Bringing a slightly higher end model into that price range might be enough to get them interested.
View-through analysis can also help in another way. Nobody ever said that when you embed a link to your website in an e-mail, it has to point to your home page. If you understand what pages users are gravitating toward, why not send them directly to what they’re interested in? Making the shopping and purchase experience convenient is a key part of making a sale.
Creating customized landing pages turns out to be a very powerful tool in your e-mail campaign tool kit. These pages are easy to create and they can be modified almost in real time based on the data you receive from your ESP. If the products you feature on your landing page don’t seem to be capturing many clicks, you can change them very quickly.
Serving Up Web Content Dynamically
Your web page is an ideal place to personalize and customize content and messages to your customers’ preferences and needs. Whether the user is logged in or not, you have a large amount of information at your disposal. This information can be used to make decisions about what content, advertisements, and links are displayed to the user.
Web analytic platforms are also an important part of your online marketing tool kit. These platforms provide detailed analysis regarding website traffic, page views, click streams, keyword searches, and everything else that leaves a data trail online. These platforms provide information not only on your site but on search engines and sites that you buy advertising on. Google Analytics and Omniture (which is now owned by Adobe) are two of the big players in this space.
Using information about browsing behavior
Chapter 10 talks about event-triggered marketing tactics. These are communications that are sent out based on the occurrence of specific events. The way you serve up content on your website is event-triggered marketing on steroids.
As soon as a customer lands on your website, everything they do can be tracked by the site. This means that every mouse click and search query can potentially be used to determine what they see next. Even if you know nothing else about the user, this browsing information is useful.
Chapter 10 also explains the notion of cross-selling. The idea is that certain products are bought in bundles or are otherwise similar in some way. So when a customer buys one product in that bundle, they are likely to buy another. One extremely common online cross-selling technique is to respond to purchases or even views of a product by immediately serving up images of similar products. You’ve no doubt experienced the “People who bought this book also bought these” pitch.
In addition to content related to your own products, your website may contain links to other websites. You may even sell advertising to other companies. You can control the links and ads that you choose to display by monitoring how the user is navigating your website.
So far, I’ve been talking about information related to a particular browsing session. There is actually a good deal more information available to you that can help you customize your web content.
Using information about the customer
The problem with data collected during a particular browsing session is that once the session ends, you lose the connection to the customer. This data is still useful for analytic purposes. But it can no longer be used to steer the conversation with the customer who created it.
Personalizing your website
If the user is registered on your website and logs in to view it, then this problem disappears. You have full visibility not only to past browsing history, but also to preferences and personal information the customer has provided.
Many websites, particularly those that have aspirations to being home pages, allow registered users to fully customize what they see. I’ve been using Yahoo! mail for years. By logging into the site, I can use MyYahoo! to essentially create my own personal web page. I can choose background colors and themes as well as page formats. I can insert links to other sites. I can choose what news stories I want displayed, and so on.
Another way to personalize your website is to tailor it around your relationship with your customer. Online banking is a classic example. When I log in to my bank’s website, I land on a homepage that is literally my own personal web page. This page contains information on all the accounts I have. It allows me to drill down to view details of individual accounts, even to the transaction level. It also allows me to open new accounts.
This last feature is no accident. The bank uses information about my accounts and how I use them to serve up ads on “my” website that it thinks I might be interested in. After we sold our last house, we paid off the mortgage we had with this bank. We went with a different lender when we bought a new house. But the bank, being astute, started to serve up ads to me regarding home equity lines of credit.
If a customer logs in to your website, it makes personalizing content easy. But there is another solution to this problem that doesn’t require the user to register or log in.
Linking web sessions together
Linking browser sessions together is done through the use of cookies. I talk more about cookies in the next section, but essentially, cookies are just small files that your website stores on a user’s computer. Cookies contain information that you want to remember and are a way of making information from one web session available when a user returns and initiates another session.
For me, like many people these days, the web has taken the place of the newspaper in my morning coffee routine. I have a news site that I check every morning. I haven’t bothered to register on the site and set up any preferences, but the site seems to know a great deal about my morning routine.
For example, during baseball season, I always check the Detroit Tigers’ box score and the Central Division standings. At the beginning of the season, I had to click the sports page and then click the MLB tab and so on to see what I wanted. Now, when I click the sports page, this information automatically appears.
The same thing is true on the financial page. I have a group of stocks that I look at every day. I no longer have to enter each ticker symbol to check the price. The ticker symbols and prices are all neatly displayed when I land on the finance home page.
This is all being done via cookies. The website has written notes to itself (on my computer) to remind it of what I have done in the past. It then serves up content accordingly.
Recognizing Customers Online
In a perfect world, every time a customer visited your website, they would log in and therefore identify themselves. But you know from your own web browsing experience that this is the exception and not the norm. In fact, the vast majority of web browsing is done anonymously, or at least without the user being logged in to your website. But unregistered browsing sessions aren’t completely anonymous. There are a number of ways of getting at least some information about users that visit your website. I briefly describe a few common techniques in this section.
Recognizing where they are: IP addresses
Whenever a user initiates a web session, the web needs to know the location of the machine requesting access in order to route content to it. This location information is known as the machine’s IP address. (IP stands for internet protocol.) When a user visits your website, their IP address is available to you. This means that (theoretically) you know where the user is located.
Location data can be quite useful in customizing web content. If a customer is shopping for pizza delivery, you may be able to use their IP address to point them to locations that are close by. If your business is selling college logo sweatshirts, you could serve up images depicting the logos of nearby colleges. In Chapter 7, I talk about a number of ways to use geographic data. These strategies apply equally well to customizing the web pages and images that you serve up to users of your website.
Location information is often used to aid in security and user verification. Whenever I’m away from home, if I try to log in to my online banking account, I have to go through an extra layer of security. I get asked one of my security questions before the site will let me in. The website recognizes that I’m not where I normally am (home) when I typically do my banking.
But wait! It isn’t quite that simple
Earlier I said that you theoretically know where your user is located. This last example points up a subtlety that you should be aware of. Actually a lot of subtleties surround the subject of IP addresses and how data is routed around the World Wide Web. The inner workings of the web are vastly complex and far beyond the scope of this book.
The particular subtlety I’m talking about is that IP addresses do not actually reflect the location of the user’s machine. They reflect the location of the device that’s actually accessing the internet. If you’re accessing the Internet from home, this means your modem. If my wife and I are both browsing the same site at the same time, both sessions will be associated with the same IP address — that of our modem.
There are a number of situations in which IP addresses can provide misleading or just plain wrong location data. One has to do with web sessions that are initiated from work rather than home. Many — in fact, most — companies make use of internal networks to connect computers, printers, and other devices. All the computers attached to these internal networks may not be in the same place. In the case of large companies, they may be spread out all over the country or even the world. But the Internet access devices that they’re using may not be nearly that spread out. Users may be remotely accessing the Internet devices, and so their actual locations may not be reflected by the IP address that shows up at your website.
IP addresses and cellular networks: Geolocation
When users access your website via mobile devices, IP addresses are particularly untrustworthy. As if directing traffic on the web weren’t complicated enough, mobile devices that use a cellphone network throw a whole other level of complexity into the mix. Mobile devices are just that. And they do move. Even when they’re stationary, the cellular network may be routing and rerouting calls through its system of towers, fiber optic lines, and various other network components. This means that the actual Internet access point may wander around during a session. Depending on traffic patterns, the actual IP address associated with a smartphone may be hundreds of miles away from the actual device.
Luckily, you have another location option for mobile devices. Virtually every mobile device is equipped with geolocation capabilities. Typically, the user has the option of turning those services off, but many people leave them on because they’re critical to a number of popular apps. Maps, searches for nearby restaurants, even keeping track of the local time while you’re traveling all depend on the mobile device knowing where it is.
As Chapter 2 explains, customer privacy is a sensitive issue. It’s particularly sensitive when it comes to collecting location data from mobile devices. The legal environment surrounding geolocation data is evolving quickly. You need to stay abreast of the legislation that’s coming down the pike. An example of such legislation has been proposed by Congressman Ed Markey of Massachusetts.
Recognizing who they are: Cookies
Chapter 3 discusses constructing a customer (and ultimately a household) record. The data you have about your customer is spread out across your enterprise, and not all of it is easily associated with individual customers. If your website is designed as an e-commerce site — that is, customers can purchase online — then many of the problems of the offline world go away. There isn’t a problem linking transaction data to the customer profile that is created when a customer registers. But identifying users, even if they are registered, can sometimes be a problem.
Using cookies to identify users
As mentioned earlier, a cookie is a file that a website deposits on a user’s machine. It’s basically the website’s way of writing itself a reminder note about something. Sometimes cookies are used to facilitate the login process on your website. When a user registers on your site, you can drop a cookie on their machine that remembers the username they registered under. Then when the user returns to your site, you can serve up a login page that has the username prepopulated. Cookies can also be used as a routine security check. If a registered user tries to log in from a machine that doesn’t contain your username cookie, you can prompt them to answer a security question to validate that they’re really who you think they are.
Earlier in this chapter, I talk a little bit about using cookies to link together information about different browsing sessions. Knowing what the user typically browses for allows you to serve up relevant content. Ideally you would like to have the user registered, because that gives you a full picture of their contact information and preferences. But there is a situation in between. For example, I can register on your site and provide you all the information you want. But if I log off and later return, I may not bother to log in.
In this situation, a cookie can be very helpful. When I come back to your site, the cookie remembers my username, even though I haven’t logged in. This allows you to access my information in the customer’s registration profile and serve up content accordingly. If I decide to make a purchase, you need to get me to log in to verify that it’s really me. But while I’m shopping, there’s nothing wrong with looking up my registration information and addressing me as Mr. Semmelroth, as opposed to dave3784.
The deleted cookie problem
Cookies aren’t a magic bullet for identifying and tracking web browsing behavior. Many users choose to delete cookies for one reason or other. Different browsers treat cookies differently. It’s also possible for users to completely block websites from depositing cookies.
Estimates of the percentage of users who delete or block cookies run as high as 40 percent. But there are some nuances to cookies that make these types of estimates a little hard to interpret. First of all, cookies come in a couple different varieties. Your online banking website probably deposits cookies on your computer to aid in security and identification and to allow you to customize certain features of your session. These cookies, deposited by the website you’re visiting, are known as first-party cookies.
Most people don’t generally block or delete these cookies. Deleting them frequently makes sites more inconvenient to use. For example, if I blocked or deleted cookies on my machine, I’d have to re-enter my location every time I checked the weather forecast rather than having it appear automatically. The stocks I follow, the baseball scores, and a host of other things that appear automatically when I visit certain sites would all have to be re-entered on every visit without these first-party cookies.
The cookies that are more frequently blocked or deleted are third-party cookies. These are cookies that are placed on your machine by websites other than the one you’re visiting — typically websites who are paying for advertising or tracking web surfing behavior. People are far more likely to block these cookies altogether. In fact, Apple’s Safari web browser blocks all third-party cookies by default.
Most of the studies that have been done regarding cookie blocking and deletion take this third-party versus first-party cookie distinction into account. These studies also have a time frame associated with them. For example, the 40 percent estimate I mention earlier actually means that 40 percent of users either block third-party cookies altogether or delete third-party cookies at least once a month. In contrast, users rarely block first-party cookies because that makes web surfing difficult.
The device problem
Fully understanding your customer’s online behavior requires knowing what they’re doing on all their devices. This is no small task. In my household alone, we have two laptop computers, three tablets, and three smartphones, not to mention a box for streaming movies from the Internet.
These devices are all connected to our private wireless network and from there to the Internet. And that’s just while we’re at home. When we’re away, eight of those devices can be connected to the internet via wireless hotspots, and six of them frequently use cellphone networks to access the web.
It’s extremely difficult for marketers to connect the myriad browsing, shopping, and purchasing behavior among all these devices and access points. There is no magic bullet for it. The most effective way of tying customer browsing behavior together is still to get them to log in to your site when they’re browsing. Short of that, you need to take advantage of the browsing information that you do have available.
Another challenge related to mobile devices is that content needs to be designed differently. It isn’t just a matter of serving up web content on a smartphone. Companies create separate, pared-down, “mobile friendly” websites for smartphones and tablets, for example. These sites typically have limited content compared to the full-blown website.
This aspect of online marketing is evolving very rapidly. A huge amount of valuable customer data is associated with these devices. It’s important for you to understand how these challenges are being approached and what solutions are emerging in the marketplace. It’s also important for you to understand how privacy law is evolving to address the concerns of consumers.
Customer Data and Search Engines
When it comes to using analytics to serve up web content, search engine sites are at the top of the heap. This is their bread and butter. Many of the browsing sessions initiated on your website come from a search engine page.
This means it’s extremely important for you to make sure that your website gets displayed prominently when potential customers are searching for information related to something you sell. Search engine optimization, or SEO, as this process is called, is a subject unto itself. Most medium to large companies have a team dedicated to doing nothing else. But your customer database plays a role.
In its simplest form, SEO involves identifying keywords you think are relevant to your product offerings. If you own a resort hotel, for example, you would want your website to appear in searches that include words and phrases like vacation and spring break in addition to searches of your company name.
In the good old days, search results were primarily based on search history. The search engine looked at where users went after doing a similar search and served up the websites in order of popularity. Search engines now take into account the customer profiles of their users. Search results are informed by an individual’s previous searches as well as other past web browsing behavior.
More and more of the screen space on the results pages of many search engines reflects sponsored search results. There is an ongoing bidding war for sponsored search engine placement. And by ongoing, I mean continuous. Your website can appear at the top of the list for a given search one day and disappear from the first page of results the next.
Search engines do a great deal of customer profiling based on search behavior. When your search optimization team is bidding for prime real estate on a results page, they have the opportunity to bid on particular customer segments as defined by the search engine profile.
Here’s where you come in. You too have a great deal of customer profile information. In Chapter 7, I talk about profiling customers and creating customer segments. The customer segments you create from your data can be mapped to search engine profiles. This allows your optimization team to focus their bidding efforts on the right customers.