4. Analytics on Web Data: The Original Big Data

Bill Franks

Wouldn’t it be great to understand customer intent instead of just customer action? Wouldn’t it be great to understand your customers’ thought processes as they decide whether they’ll make a purchase? In the past it was virtually impossible to get answers to such questions. Today, they can be answered with the use of detailed web data. That’s what this chapter is all about.

There is no better way to understand what big data is all about than to see some specific examples of big data and how it can be used. Perhaps no big-data source is as widely used today as web data.

Organizations across a number of industries have integrated detailed, customer-level behavioral data sourced from websites into their enterprise analytics environments. Most organizations, however, still end web integration with the inclusion of online transactions. Traditional web analytics vendors provide operational reporting on click-through rates, traffic sources, and metrics based only on web data. However, detailed web behavior data historically was not leveraged outside of web reporting.

Leading companies have shown that detailed web data can provide previously untapped corporate value. This chapter outlines what those leaders are doing, why they are doing it, and why every organization should consider such analytics today. The examples are compelling and promise to be eye-opening to those who have never given much thought to integrating detailed clickstream data with other data as opposed to keeping it isolated.

The core theme of this chapter isn’t simply the taming of web data. Instead of aggregated web metrics from a distinct data silo, organizations should focus on integrating web data with all the other relevant information about their customers. Utilizing such information in a scalable analytics environment lets you move beyond purchasing insights about customers and into individual intentions, purchase decision processes, and preferences. Tapping into the rich insight provided by this new data source, an organization can make huge strides forward.

How does an organization capture, analyze, and utilize this rich information to drive insight? First, we’ll discuss what data needs to be acquired and why. Then we’ll explore some examples of what that data can reveal. Finally, we’ll discuss specific examples of how analytics processes can be transformed through the integration of web data. Web data is one big-data source that many organizations have already tamed. Add yours to the list!

Web Data Overview

Organizations have talked about having a 360-degree view of their customers for years. At any point in time, one organization or another claims that it has achieved a true 360-degree view. In reality, it is impossible to have a true 360-degree view because this implies that you know everything there is to know about your customers. When a 360-degree view is discussed, what is really meant is that the organization has as full a view of its customers as possible considering the technology and data available at that point in time. However, the finish line is always moving. Just when you think you have finally arrived, the finish line moves farther out again.

A few decades ago, companies were at the top of their game if they had the names and addresses of their customers and could append demographic information to those names through the then-new third-party data enhancement services. Eventually, cutting-edge companies started to attach basic recency, frequency, and monetary (RFM) value metrics to customers. Such metrics look at when a customer last purchased (recency), how often she has purchased (frequency), and how much she spent (monetary value). These RFM summaries might be tallied for the past year and possibly over the customer’s lifetime. In the past 10 to 15 years, virtually all businesses have started to collect and analyze their customers’ detailed transaction histories. This had led to an explosion of analytical power and a much deeper understanding of customer behavior.

Many organizations are frozen at the transactional history stage. Although this stage is still important, many companies incorrectly assume that this is still the closest they will get to a 360-degree view of their customers. Today, organizations need to collect newly evolving big-data sources related to their customers from a variety of extended and newly emerging touchpoints such as web browsers, mobile applications, kiosks, social media sites, and more.

Just as transactional data enabled a revolution in the power and depth of analysis, so too do these new data sources let you take analytics to a new level. With today’s data storage and processing capabilities, it is absolutely possible to achieve success, and many forward-thinking companies have already proven it by applying the data to a variety of problems, some of which we’ll discuss shortly.

What Are You Missing?

Have you ever stopped to think about what happens if only the transactions generated by a website are captured? Perhaps for a website, 95% of browsing sessions do not result in the creation of a shopping basket. Of the 5% that do, only about half, or 2.5%, actually begin the checkout process. And of that 2.5%, only two-thirds, or 1.7%, actually complete a purchase. These figures are not unrealistic in many cases.

What this means is that information is missing on more than 98% of web sessions if only transactions are tracked. But more important, an even higher percentage of available data is missing. For every purchase transaction, dozens or hundreds of specific actions might be taken on the site to get to that sale. That information needs to be collected and analyzed alongside the final sales data.

It is important to note that this is not just the same old web analytics story from years past. Traditional web analytics focus on aggregated behavior, summarized in an environment where only web data was included. The goal needs to be moving beyond reporting summary statistics, even if they can be viewed in some detail, to actually combining customer-level web behavior data with other cross-channel customer data. This is moving far beyond clickthrough reports and page view summaries.

Just as RFM is only a small piece of what transaction data can yield, so too are traditional web analytics only a portion of what web data can yield. Web data is a game-changing, amazing new frontier that can revolutionize organizations’ customer insights and the impacts those insights have on their businesses.

Imagine the Possibilities

Imagine knowing everything customers do as they go through the process of doing business with your organization. Not just what they buy, but what they are thinking about buying, along with the key decision criteria they use. Such knowledge enables a new level of understanding about your customers and a new level of interaction. It allows you to meet their needs more quickly and keep them satisfied.

• Imagine you are a retailer. Imagine walking through the aisles with customers and recording every place they go, every item they look at, every item they pick up, every item they put in the cart and then take back out. Imagine knowing whether they read nutritional information, if they look at laundry instructions, if they read the promotional brochure on the shelf, or if they look at other information made available to them in the store.

• Imagine you are a bank. Imagine being able to identify every credit card option customers consider. Imagine being able to understand if it was a reward program, interest rates, or annual fees that drove their choice. Imagine knowing what they say about each product after they own it.

• Imagine you are an airline. Imagine being able to identify every flight customers view before choosing their final itinerary. Imagine knowing if they care more about price or convenience. Imagine knowing all the destinations they consider and when they first consider them.

• Imagine you are a telecom company. Imagine being able to identify every phone model, rate plan, data plan, and accessory customers consider before making a final decision. Imagine knowing that they came back to your site by typing into a search engine “renew contract” or “contract cancellation.”

It certainly sounds exciting to have the information outlined in this list. You can have it right now by making a commitment to collect it and make it available for analytics. Organizations in each of these industries are already doing so.

A Fundamentally New Source of Information

The beauty of exploring customers’ detailed web behavior is that it moves beyond just knowing what they buy. You can now gain insights into how they make their decisions. Instead of seeing just the result, you have visibility into the entire buying process. This big-data source isn’t a simple extension of existing data sources. Many organizations were excited about the integration of web transactions with traditional transactions. But, at its core, a web transaction is simply another transaction record with a new “transaction type” or “transaction location” flag. In the case of detailed web behavior, there is no existing analog for most of the data that can be collected. It is a fundamentally new source of information.

One of the most exciting aspects of web data is that it provides factual information on customer preferences, future intentions, and motivations that are virtually impossible to get from other sources outside of a direct conversation or survey. Why do customers choose one offering over another? Perhaps organizations think they know. However, they will likely find that many customers make choices in ways that were not anticipated.

As soon as you know customers’ intentions, preferences, and motivations, you have completely new ways of communicating with them, driving further business, and increasing their loyalty. The glorious part of this story happens when you marry web data with everything you have learned from the prior 360-degree view. Now you can extend that view with all the rich new web behavior data available.

What Data Should Be Collected?

If possible, you should capture any action that a customer takes while interacting with an organization. That means detailed event history from any customer touchpoint. Common touchpoints today include websites, kiosks, mobile apps, and social media. You can capture a wide range of customer events:

• Purchases

• Product views

• Shopping basket additions

• Watching a video

• Accessing a download

• Reading/writing a review

• Requesting help

• Forwarding a link

• Posting a comment

• Registering for a webinar

• Executing a search

This chapter focuses on web data, but the same principles apply for the other sources listed. The examples that follow are websitecentric, but keep in mind that the same concepts apply across the board to all touchpoints from which data can be collected.

What About Privacy?

Privacy is a big issue today and may become an even bigger issue as time passes. You must seriously consider what data is captured and how it is used. You need to respect not just formal legal restrictions, but also what your customers will view as appropriate. The last thing an organization wants to do is create programs that customers view as being “creepy” or intrusive. Privacy is an issue worthy of a deep discussion within your organization. It is outside the scope of this chapter to cover all the issues surrounding privacy. However, we will examine one option to address privacy concerns while still gaining value from analyzing web data.

Even if an organization wants to be conservative in its actions, there are options for realizing tremendous value from web data. Even if you have no desire to interact with customers individually or tie the data back to identifiable customer data, web data is still valuable. An arbitrary identification number that is not personally identifiable can be matched to each unique customer based on a logon, a cookie, or a similar piece of information. This creates what might be called a “faceless” customer record. Even though all the data associated with one of these identifiers is from one person, the people doing the analysis have no ability to tie the ID back to the actual customer. Analysis can still be done to look for patterns across customers, however. These patterns are powerful and can be found without ever worrying about which specific individual did what.

It is the patterns across faceless customers that matter, not the behavior of any specific customer. The individuals in this example are important only as an input to the pattern analysis. Nobody needs to identify who each individual actually is to derive value. With today’s database technologies, analysts can perform analysis without being able to identify the individuals involved. This can remove many privacy concerns. Of course, many organizations do in fact identify and target specific customers as a result of such analytics. They have presumably put in place privacy policies, including opt-out options, and are careful to follow them.

What Web Data Reveals

Now that we’ve covered what web data is, let’s dive into it in more detail. There are a number of specific areas where web data can help organizations understand their customers better than is possible without web data. Without taming this source of big data, such insights will be very difficult, if not impossible, to come by. We’ll establish some broad categories of the kinds of insights you can gain from web data in this section before moving on to detailed use cases and applications in the final section.

Shopping Behaviors

A good starting point for understanding shopping behavior is identifying how customers come to a site to begin shopping. What search engine do they use? What specific search terms do they enter? Do they use a bookmark they created previously? Analysts can take this information and look for patterns in terms of which search terms, search engines, and referring sites are associated with higher sales rates. Note that analysts can look into higher sales rates not just within a given web session, but also for the same customer over time. This can be combined with a view of sales on the website along with a cross-channel view of the customer’s purchase behavior over time. That is where the value resides.

As soon as customers are on a site, start to examine all the products they explore. Identify who simply looked at a product landing page and left, and who drilled down farther. Who viewed extra photos? Who read product reviews? Who looked at detailed product specifications? Who looked at shipping information? Who took advantage of any other information that is available on the site? For example, identify which products were chosen for a “Compare” view. Last, it is easy to identify which products were added to a wish list or basket, as well as if they were later removed.

One interesting ability enabled by web data is identifying product bundles that are of interest to a customer before she makes a purchase. Move beyond trying to up-sell a customer with a follow-up offer after a purchase. Instead, examine what she is browsing and make her an offer to buy a complete bundle in the first place.

For example, consider a customer who views computers, backup disks, printers, and monitors. It is likely the customer is considering a complete PC system upgrade. Offer a package right away that contains the specific mix of items the customer has browsed. Do not wait until the customer purchases the computer and then offer generic bundles of accessories. A customized bundle offer before the customer buys is more powerful than a generic one after she has purchased.

Customer Purchase Paths and Preferences

Using web data, analysts can identify how customers arrive at their buying decisions by watching how they navigate a site. It is possible to gain insight into their preferences. Consider an airline, which can tell a number of things about preferences based on the ticket that is booked. For example, how far in advance was the ticket booked? What fare class was booked? Did the trip span a weekend? This is all useful information, but an airline can get even more from web data.

An airline can identify customers who value convenience. Such customers typically start searches with specific times and direct flights only. They deviate from the most convenient direct flight only if there is a huge price difference for a minimal change in convenience. Perhaps a customer can save $700 by flying into New York’s JFK airport instead of LaGuardia. He can land at JFK within 30 minutes of the LaGuardia flight, and the extra cab fare is only about $20. In that case, a convenience-oriented customer might decide $700 in savings is worth the extra hassle of JFK. But if the difference is only $200 and the arrival time is two hours later, a convenience-oriented customer will stick with the more convenient option.

Airlines can also identify customers who value price first and foremost and are willing to consider many flight options to get the best price. Such customers will deviate from the cheapest option only if there is a moderate price difference for a huge gain in convenience. For example, perhaps a customer can leave at 10 a.m. for $220 versus leaving at 6 a.m. for $200. The four extra hours of sleep are worth $20 to a price-oriented customer, so she pays the $20 premium for the later flight.

Based on search patterns, airlines also can tell how tied to deals or specific destinations a given customer is. Does she research all the special deals available and then choose one of those for her trip? Or does she look only at a certain destination and pay what is required to get there? For example, a college student may be open to any number of spring break destinations and will take the one with the best deal. On the other hand, a customer who regularly visits family will only be interested in flying to where her family is.

Simply knowing that a customer regularly browses weekend deals for certain destinations can be a good indicator of what’s important to her. Some customers are open to visiting family whenever they see a deal to the right city. If they see a deal, they book it. Once that pattern is identified, an airline can anticipate customers’ needs better.

In the preceding examples, historical insight into purchase history is invaluable when married with current browsing and research patterns. Of course, it takes time and effort to change analytical processes to account for such patterns. But as soon as you know which aspects of a site appeal to customers on an individual basis, you can target them with messages that meet their needs much more effectively.

Research Behaviors

Understanding how customers use a site’s research content can lead to tremendous insights into how to interact with each customer. You also can figure out how different aspects of the site do or do not add value in driving sales. As you examine the options customers explore on their way to a purchase, you can infer what is important to them.

For example, consider an online store that sells movies. If some customers routinely look at the standard, widescreen, extended, and HD versions of a movie before making a final decision, that says they are open to various format options even if they often end up buying a certain format most of the time. As soon as you know a customer’s patterns, you can alter what she sees when she visits a site to make it easier for her to find her favorite options quickly. A customer who views a lot of formats might be shown all the formats every time. However, why make a customer sort through all DVD formats if you know that she neither browses nor buys anything but a single format?

Another way to use web data to understand customers’ research patterns is to identify which pieces of information offered on a site are valued by the customer base overall and by the best customers specifically. How often do customers look at previews, additional photos, or technical specs before making a purchase? Note that when you track across sessions and combine with other customer data, you can know if people researched one day and then bought another day. A final purchase event often is a highly targeted web session that simply executes the purchase. The historical browsing history is needed to put together the whole picture. Perhaps a little-used website feature the organization was considering removing is a big favorite among a critical segment of customers. In that case, the feature might be kept.

An organization might see an unusual number of customers drop a specific product after looking at the detailed specifications page, but not when they don’t view the specs. After looking into what is on the page, the company might find that the product description is unclear or that one of the specs is inaccurate. With an updated description, sales increase.

The reading of reviews is a tremendous indicator of what is important to people. Which customers value reviews? Which do not? Which products routinely lose customers after their reviews are read? Reviews have the power to make or break a sale. Once you know which customers usually buy after reading reviews, if you see many of them deciding not to purchase a specific product after reading its reviews, you should look into it. Perhaps some negative reviews are posted. If so, you can identify if they are valid, what points they raise, and how you will address those points.

In the end, identifying which site features are important to each customer and how each customer leverages the site for research can help you better tailor a site to the individual. For customers who always drill down to detailed product specifications, perhaps those specs can come up as soon as a product is viewed. For those who always want to see photos, perhaps photos can be featured in full size instead of as thumbnails. The point is to make research easier for your customers so that they will come to you instead of the competition when they are ready to research and buy.

Feedback Behaviors

Some of the best information customers can provide is detailed feedback on products and services. Simply the fact that customers are willing to take the time to offer feedback indicates that they are engaged with a brand. By using text mining to understand the tone, intent, and topic of a customer’s feedback, you begin to get a better picture of what is important to that individual.

Do certain customers regularly post reviews of what they buy? If those reviews are often positive and are read by other customers, it might be a good idea to give such customers special incentives to keep the good words coming. Similarly, if you study the questions and comments submitted via online help chats with customers, you can get a feel not just for what customers in general are asking about, but what each specific customer is asking about. If analysis shows that certain features are always important for a specific customer, point the customer in the direction of other items with similar attributes.

Is a customer a fan of your company on Facebook? Does he or she follow you on Twitter? By looking at the comments and questions customers pose through such interfaces, you can learn much about their likes and dislikes. Additionally, when you identify very active customers who often write about your company on social media sites, you may want to cultivate them as an influential brand ambassador. Give such customers the extra attention they deserve given the influence they have over your brand. Note that customers’ influence is not always strongly correlated with their individual value. A midsized customer who usually warrants standard treatment can be very vocal. It may be smart to upgrade him beyond what his dollar value implies due to the influence he wields.

Web Data in Action

What an organization knows about its customers is never the complete picture. You must always make assumptions based on the information available. If you have only a partial view, you often can extrapolate the full view accurately enough to get the job done. But it is also possible that the missing information paints a totally different picture than you expected. In the cases where the missing information differs from the assumptions, you can make suboptimal, if not totally wrong, decisions.

Therefore, organizations should strive to collect and analyze as much data as possible. We’ve discussed a number of different types of web data and some broad uses of them. Now, let’s move on to some specific examples of how organizations can apply web data to enhance existing analytics, enable new analytics, and improve their business.

The Next Best Offer

A very common marketing analysis is predicting the next best offer for each customer. Of all the available options, which single offer should next be suggested to a customer to maximize the chances of success? Having web behavior data can totally change the decision of what a customer’s next best offer is and make those decisions much more robust.

Let’s assume that you work at a bank and that you know the following information about a customer named Mr. Smith:

• He has four accounts: checking, savings, credit card, and a car loan.

• He makes five deposits and 25 withdrawals per month.

• He never visits a branch in person.

• He has a total of $50,000 in assets deposited.

• He owes a total of $15,000 between his credit card and car loan.

What is the best offer you could email Mr. Smith? Based on his profile, it would be reasonable to argue for any number of things such as a lower credit card interest rate or an offer of a certificate of deposit for his sizable cash holdings. One thing that would not be high on most people’s list is offering a mortgage because no data says this choice is remotely relevant. However, when you examine Mr. Smith’s web behavior, a couple of key facts jump off the page:

• He browsed mortgage rates five times in the past month.

• He viewed information about homeowners’ insurance.

• He viewed information about flood insurance.

• He explored home-loan options (fixed-rate versus variable-rate, 15-year versus 30-year) twice in the past month.

It’s pretty easy to decide what to discuss next with Mr. Smith now, isn’t it?

It can be difficult for any business to determine if its customer base is still engaged. The web provides direct clues about what interests customers and if they are still engaged. Consider the case of a catalog retailer that also has many store locations. The cataloger collects the following data for each customer, among other things:

• Last products browsed

• Last products reviewed

• Historical purchases

• Marketing campaign and response history

The data is compiled and analyzed to determine which products each customer appears most interested in. Adjustments are made to the content of catalogs sent, as well as the length of the catalogs and the offers within each. The effort leads to major changes in the cataloger’s promotional efforts versus its traditional approach, providing the following results:

• A decrease in total mailings

• A reduction in total catalog promotions pages

• A materially significant increase in total revenues

Web data can help completely overhaul activities for the better.

Attrition Modeling

In the telecommunications industry, companies have invested massive amounts of time and effort to create, enhance, and perfect churn models. Churn models flag customers who are most at risk of canceling their accounts so that action can be taken proactively to prevent them from doing so. Churn is a major issue for the industry; huge amounts of money are at stake. The models have a major impact on the bottom line.

Managing customer churn has been, and remains, critical to understanding patterns of customer usage and profitability. Imagine how this has been invigorated today with the use of web data put into the right context. Mrs. Smith, as a customer of telecom Provider 101, goes to Google and types “How do I cancel my Provider 101 contract?” She then follows a link to Provider 101’s cancelation policies page. Imagine how much stronger, more time-sensitive, and usable this customer data is for a churn model and for taking meaningful action compared to other data.

It is hard to think of an indicator of cancelation that is stronger than knowing that Mrs. Smith researched canceling—aside from her actually taking the final step of making the cancelation request. Perhaps analysts would have seen her usage dropping, or perhaps not. It would take weeks or months to identify such a change in usage. By capturing Mrs. Smith’s actions on the Web, Provider 101 can move more quickly to avert losing her as a customer.

Missing early opportunities to identify customers who are exploring cancelation means trying to win them back when their minds are already made up and another carrier may have already won their business. It will be too late in most cases, and the customer will be lost for good.

Response Modeling

Many models are created to help predict the choice a customer will make when presented with a request for action. Models typically try to predict which customers will make a purchase, or accept an offer, or click a link in an email message. For such models, a technique called logistic regression is often used. These models usually are called response models or propensity models. The attrition model we discussed a moment ago is in the same class of model. The main difference is that in an attrition model, the goal is predicting a negative behavior (churn) rather than a positive behavior (purchase or response).

When using a response or propensity model, all customers are scored and ranked according to their likelihood of taking action. Then appropriate segments are created based on those rankings to reach out to the customers. In theory, every customer has a unique score. In practice, because only a small number of variables define most models, many customers end up with identical or nearly identical scores. This is particularly true of customers who do not spend much or frequently. In such cases, many customers can end up in big groups with very similar, very low scores.

Web data can greatly increase differentiation among customers. This is especially true among low-value or infrequent customers, who can have a large uplift in score based on the web data. Let’s look at an example where four customers are scored by a response model with a handful of variables. Each customer in the example has the same score due to having the same value for each of the model’s variables. The scores are hypothetical, so don’t worry about how they were computed. The four customers’ profiles are as follows:

• Last purchase was within 90 days

• Six purchases in the past year

• Spent $200 to $300 total

• Homeowner with estimated household income of $100,000 to $150,000

• Member of the loyalty program

• Has purchased the featured product category in the past year

In this case, all customers get the exact same score and look identical in terms of their likelihood to respond. Let’s assume they all score 0.62. Any marketing program based on this model will treat each of these four customers the same. After all, based on the preceding information, nothing differentiates them; they are exactly the same!

Now, using web data, let’s see how drastically the view changes. Look how the web data provides powerful new information:

• Customer 1 has never browsed your site, so his score drops to 0.54.

• Customer 2 viewed the product category featured in the offer within the past month, so her score rises to 0.67.

• Customer 3 viewed the specific product featured in the offer within the past month, so his score rises to 0.78.

• Customer 4 browsed the specific product featured three times last week, added it to a basket once, abandoned the basket, and then viewed the product again later. Her score rises to 0.86.

This web behavior allows us to identify customers with a current interest, if not intention, to purchase. It is possible to score customers better and end up with solid differentiation among them, where originally there was none. Now, repeat the example of these four customers across millions of customers across multiple channels, and dramatic changes can be driven!

When asked about the value of incorporating web data, a director of marketing from a multichannel American specialty retailer replied, “It’s like printing money!” The good news is that it is very easy to build a model both with and without web data to prove exactly how results improve for any given situation. There is virtually no risk in testing the impact in your organization’s environment.

Customer Segmentation

Web data also enables a variety of completely new analytics. One of those is to segment customers based solely on their typical browsing patterns. Such segmentation provides a completely different view of customers than traditional demographic or sales-based segmentation schemas. In addition, such segmentation can yield unique insights and actions.

Consider a segment called Dreamers that has been derived purely from browsing behavior. Dreamers repeatedly put an item in their baskets but then abandon them. Dreamers often add and abandon the same item many times. This may be especially true of high-value items such as TVs and computers. It isn’t difficult to clearly identify the segment of people who do this repeatedly. So what can you do after finding them?

One option is to look at what the customers are abandoning. Perhaps a customer is looking at a high-end TV that is quite expensive. You’ve seen in the past that this customer often aims too high and eventually buys a less-expensive product than the one she abandoned repeatedly. Sending her an email pointing out less-expensive products that have many of the same features may be a way to get her to buy a TV sooner.

Another option is operational in nature. Abandoned-basket statistics can be adjusted to account for the Dreamer segment. Organizations often view abandoned baskets as a failure. However, by examining the browsing history, you find that 10 abandons were from one customer who is known to repeatedly and regularly abandon many products. As a result, the abandoned-basket count can be reduced, and all the customer’s abandons for that product can be counted as a single abandonment. This yields a cleaner view of abandonment. By the time statistics are adjusted for all such customers, the average abandonment rate might look quite a bit better. Not only do the new figures look better, but they also are a more accurate reflection of reality.

Assessing Advertising Results

Better assessing paid search and online advertising results is another high-impact analysis enabled with customer-level web behavior data. Traditional web analytics provide high-level summaries such as total clicks, number of searches, cost per click or impression, keywords leading to the most clicks, and page position statistics. However, these metrics are at an aggregate level and are rolled up only from the individual browsing session level. The context is limited solely to the web channel.

This means that all statistics are based only on what happened during the single session generated from the search or ad click. When a customer leaves the website and his web session ends, the scope of the analysis is complete. There is no attempt to account for past or future visits in the statistics. By incorporating customers’ browsing data and extending the view to other channels as well, it is possible to assess search and advertising results at a much deeper level.

• Were the site visits each ad or search term generated associated with the most valuable or least valuable customers?

• How many sales did the initial session lead to in the days or weeks that followed the customer’s first click?

• Are certain referring websites drawing visitors who return for more visits and make more total purchases than visitors referred from other sites?

• By doing a cross-channel analysis that accounts for activity in other channels, are a lot of sales closed in a second channel after interest is generated on the Web via an ad or search?

Let’s consider an example from a financial institution. Credit card applications are everywhere. They are in the mail, they are in magazines, and they are available all over the Web. The bank in our example understands that “eyeballs and clicks” are only a portion of the picture. What happens after the initial click is the telling information about the value of an advertising placement.

The bank performs extensive analytics to dive deeper and look at more than just clicks from the initial session. Customers are examined across time and sessions to also assess application completion, customer service inquiries, card issuance, card activation, and initial credit spending. This view of advertising beyond the click provides a more complete view of advertising success and leads to smarter allocation of advertising budgets.

Through detailed, customer-level web data, you can understand which ads, keywords, or referring sites generate the “best” clicks based on a much larger picture than simply aggregated results from initial web sessions. With the additional insight provided by the extended cross-channel, cross-time view, you can see a picture that has previously been unavailable. Organizations that understand the deeper context will have an opportunity to take advantage of new strategies that those using traditional levels of analysis will be unable to identify. That is a distinct competitive advantage.

Wrap-Up

Here are the most important lessons to take away from this chapter:

• The integration of detailed, customer-level web behavior data can transform what organizations understand about their customers.

• Just as transactional data enabled a revolution in the power and depth of analysis when it became available, so will web data allow you to take analytics to a new level.

• Other customer touchpoints can be tracked in a similar fashion as a website, such as kiosks and mobile phone applications. The same principles apply.

• Any data that can be captured should be. This includes page views, searches, downloads, and any other activity on a website.

• Privacy is a major concern with web data, so you should be careful when defining policies on how such data will be used. Those policies must be rigorously followed and enforced.

• You can generate tremendous value by analyzing faceless customers who are identified only by an arbitrary identification number. This way, neither analysts nor anyone else can identify who each customer actually is. Only the patterns matter.

• Web data helps you understand detailed customer shopping, research, and feedback behaviors and purchase paths. It is almost as if you can read your customers’ minds.

• Web data enables stronger results in areas such as next best offer, attrition modeling, response modeling, customer segmentation, paid search, and online advertising analysis.

• The opportunity to be an early adopter and get ahead of the competition is almost over. Get started taming this big-data source now!