Where Do You Start? 27 few can articulate what data-driven marketing might look like. I also found that there were consistent roadblocks that were holding marketers and organizations back from adopting data-driven marketing principles. So why is marketing measurement and data-driven marketing so difﬁcult? I have asked this question of many groups of marketing managers and executives. Following are the answers that I often hear grouped into ﬁve major categories: Obstacle 1: Getting Started We don’t know how. We don’t have the right metrics. The problem is not too little data; quite the opposite—we have lots of data, but none of the data are useful. We don’t know where to start. Obstacle 2: Causality There are too many confounding factors; overlapping campaigns make deﬁning cause and effect impossible. There is a time delay between the marketing campaign and cus- tomer action. Awareness campaigns do not directly result in sales, but our chief ﬁnancial ofﬁcer (CFO) wants to see the ﬁnancial ROI. Obstacle 3: Lack of Data We are a business-to-business (B2B) company and sell indirectly. As a result, we don’t know who our customers are. We can’t collect customer data due to privacy issues. Obstacle 4: Resources and Tools We don’t have the time and/or it cost too much. We don’t have the tools and/or systems to support data-driven marketing. We are marketers and can’t communicate with information tech- nology (IT) people. IT builds systems, but they are not the resources and tools we need.
28 Data-Driven Marketing Obstacle 5: People and Change We don’t measure because we don’t want accountability. Our incentives are all for marketing activity, not results. We do not have a culture of measurement. We don’t have the skills for data-driven marketing. Our organization is resistant to new ideas, such as data-driven marketing. Marketing is creative: imposing metrics and process will kill cre- ativity and innovation. This chapter focuses on answering the question ‘‘Where do I start?’’ and provides strategies for how to overcome these ﬁve obstacles. Overcome Obstacle 1: Getting Started—Focus on Collecting the Right Data and Create Momentum by Scoring an Easy Win Royal Bank of Canada (RBC) started the journey toward data-driven marketing by looking internally. Said Cathy Burrows, who led the RBC initiative, ‘‘You have to look at what you do today and think, ‘How can I do this better, cheaper, faster, and smarter?’ ’’ RBC started by looking at the Canadian equivalent of individual retirement account (IRA) contributions, tax-free voluntary retirement investments made once a year. Each year around IRA contribution time the sales force was given a list by the RBC marketing department. Sales then called the list, which was ordered alphabetically. For each salesperson, the average yield was one to two people accepting the offer from the list of the top 10 candidates for the salesperson to call. The RBC marketing team built a model to rank and score the list based on propensity to contribute more than $5,000 to an IRA (see Chapter 9 for how to do this). The model involved analyzing more than 1 million customers, for 12 months of data, and scoring them to ﬁnd the top 250,000 that would potentially contribute. The size of this data set may seem intimidating, but realize that if you can solve a problem for 10 customers, you can solve it for 1,000 customers; if you can solve the problem for 1,000 customers, you can do it for a million or more. (Note, however, that while the principles
Where Do You Start? 29 are the same for large customer bases, most likely you can’t use a personal computer. Chapter 10 answers the question ‘‘What’s it going to take?’’ from an infrastructure perspective when you have a large customer base.) At RBC, collecting the relevant data to do the scoring required a lot of manual legwork and took six months of effort at a cost of approxi- mately $100,000. The output of the initiative was to give each sales- person a new targeted list of the 25 top customers to call. The results were impressive: 8 out of 10 customers called by each salesperson now accepted the offer to set up an IRA. However, it took time for the sales force to realize the value of the new list. In the ﬁrst year, only 25 percent of the sales force participated, but by the third year, more than 75 percent were participating. This illustrates how new experiments with data-driven marketing cannot just be pushed down from corporate. Burrows told me: ‘‘I wanted the salespeople to say, ‘Damn, that list was good,’ and drive adoption from the bottom up. The small IRA marketing initiative was the starting point. This win enabled us to build the business case and gain executive sponsorship for the much bigger $4 million marketing initiative.’’ The point is that you don’t need 100 percent of the data and a multimillion-dollar infrastructure to get started. The key is to focus on collecting the right data. Ask what are the 20 percent of data that will give 80 percent of impact? Start there. Then show the quick win in order to get executive support and secure funding for the next stage. As a second example, Figure 2.1 is a picture of three stores of the Walgreens pharmacy chain on a map. Walgreens is a $59 billion annual revenue pharmacy company with 6,850 stores throughout the United States. This geospatial picture shows dots that are the customers and where they live and are coded by shape depending on which of the three Walgreens stores they shop. The ‘‘diamond’’ customers shop at Store 1; the ‘‘square’’ customers, at Store 2; and the ‘‘star’’ customers, at Store 3. This pharmacy retail chain predominantly markets using ﬂyers in newspapers. The way they pay for the marketing is by zip code, denoted by the dashed line, for example, in the picture. Mike Feldner, the marketing manager who ﬁrst created these pictures, noticed something interesting: the circle on the picture is two miles in radius, and after looking at many pictures throughout the United States, he noticed that there are no dots (customers) for a store more than two miles from the
30 Data-Driven Marketing Figure 2.1 A map with three Walgreens stores. The dots on the pictures are where the customers live. The customers are coded depending on where they shop. Source: Adapted from original data by Walgreens Marketing. store. He concluded that in the United States, if you live more than two miles from a pharmacy store, you probably don’t shop there. At that time, Walgreens treated each U.S. locale equally; allocating equal dollar amounts for newspaper advertising in each zip code across the United States. But the data show that if there is no store within two miles of the zip code, customers do not shop at the store. Based on these data, Walgreens ultimately stopped spending advertising dollars in all zip codes without a store within two miles of the zip code. As you might guess, the impact to sales revenues was exactly zero. The impact to marketing, however, was a cost saving of more than $5 million, for a total cost of collecting the data and creating the plots of approximately $200,000. This multimillion-dollar saving in marketing did not require a lot of money, and the analysis was done on a personal computer (PC). Walgreens already had the ESRI map and graphing software (www .ESRI.com) to manage its store locations. Feldner’s innovation was to
Where Do You Start? 31 add in Walgreens’ newspaper marketing spending. Feldner told me, ‘‘We started with simple Excel spreadsheets that contained the adver- tising circulation distributions by zip code. It was not hard to get the advertising data into the software and to create maps with the store and customer locations—we did it all on a PC.’’ The bigger challenge, however, was in changing Walgreens’ busi- ness process to use data-driven marketing insights. Feldner said to me, ‘‘When we ﬁrst started, we made too many changes too quickly and did not do a good job communicating what we were doing.’’ As a result, the Store Operations area was uncomfortable with the changes to how the advertising dollars were spent, and within a few weeks the changes were rolled back to marketing as usual. ‘‘I realized we had to start small,’’ Feldner told me. ‘‘I found a Store Operations VP and district manager who were willing to try things. I remember ﬂying with my advertising counterpart to meet with them. We showed them the pictures, and they could see that it did not make sense to spend $80,000 in advertising in a zip code ﬁve miles from the store that was generating, at most, $20,000 in sales. With ﬁve examples, we cut $300,000 off the bat and showed the approach really worked.’’ Given this win, they then set up a process to review the marketing spending with each district manager throughout the United States. The Store Operations group came around pretty quickly when they were shown how to cut their speciﬁc marketing budgets and increase proﬁt- ability. Feldner told me, ‘‘The initial failure, and then the success that came from starting over with a small win, was the biggest learning experience of my career.’’ In the mid-1990s, Continental Airlines was positively the worst airline in the industry, and this was backed up by data; Continental ranked dead last on every conceivable airline metric. The crowning moment was when David Letterman called out the airline in his nightly Top 10 list. At the height of the baseball player’s strike in the United States in 1995, the number 10 item on Letterman’s tongue-in-cheek list of baseball players’ demands was ‘‘No team ﬂights on Continental Airlines.’’ This was seen by more than 5 million viewers that night. Gordon Bethune, CEO of Continental from 1994 until his retire- ment in 2004, orchestrated an amazing worst-to-ﬁrst turnaround.3 He ﬁrst focused on making Continental clean, safe, and reliable and changed the incentive scheme for employees. Employees received $100/month cash bonus if the planes ran on time. The results were amazing—in one month Continental became the number one on-time
32 Data-Driven Marketing airline. Bethune then took his worst-to-ﬁrst approach to the next level with the goal of becoming ﬁrst-to-favorite for customers. Kelly Cook, who worked on the Continental data-driven marketing component of the initiative, told me: At Continental, the essential ﬁrst step was to talk to our customers. Focus groups were a low-cost and relatively quick way for us to test our ideas. We also gained invaluable feedback on what we as a company needed to change and what the priorities should be. We did not have the budget or the resources to integrate all of our data, so we focused on the two databases of the 45 silos that, if we combined the data, gave us the greatest return. The marketing managers had a hunch that if they sent a letter to high-value customers who had experienced a catastrophic event, de- ﬁned as lost luggage, a canceled ﬂight, or a major ﬂight delay, good things would happen. They did not have millions of dollars of consoli- dated infrastructure. They had 45 disconnected databases scattered across the company, and the marketing database had been outsourced to an agency. Their ﬁrst step was to ﬁnd the two most important databases: customer ﬂight proﬁtability and service exceptions. They then designed a marketing campaign that sent a letter within 12 hours of an event and arranged for focus groups of customers who received the letters. The letter was very simple, and said, ‘‘ . . . you are a valued customer to Continental Airlines and we are sorry. . . . ’’ As part of the experiment, some of the letters had one-pass frequent ﬂyer miles, and others had free passes to the Presidents Club, the executive club lounges in the major airports Continental served. The focus groups’ results that measured the impact of the letters were always the same. In control groups that did not get the letter, one person would explain their story of misery of ﬂight cancellation or lost luggage. Another person would then explain that he or she could top that story and give an even worse account. Before long, the group would degenerate into a hate-fest for Continental Airlines. Groups that got the letters were completely different. Representa- tive reactions were, ‘‘Wow, I had not completed my trip yet, and yet the letter was there when I returned home’’ and ‘‘No company had ever said that they were sorry to me before.’’ Their reaction was completely different. These focus groups provided qualitative evidence that there was a signiﬁcant customer perception change by sending the letter.
Where Do You Start? 33 Interestingly, a signiﬁcant percentage of those customers who received the free passes to the Presidents Club signed up for the club. This made the marketing initiative have a very positive return on marketing investment (ROMI), since the Presidents Club is a high- margin, high-proﬁtability offering for an airline. Later, when the systems were in place to integrate the 45 databases and enable the calculation of true proﬁtability and customer lifetime value (CLTV; see Chapter 6 for the deﬁnition and calculation), they found that those who received the letter had an 8 percent improvement in revenues. The takeaways are that to start the journey toward data-driven marketing, Continental started small and used focus groups and experi- ments. The early results built momentum for the marketing initiative and helped the line managers secure executive support. Mike Gorman, senior director of customer management at Continental, told me: ‘‘Creating the database was hard work. But once we had the capability, we were able to quickly spin off tools for customers and the enterprise that created an enormous amount of value.’’ Overcome Obstacle 2: Causality—Conduct Small Experiments Another objection to data-driven marketing I hear frequently is that ‘‘there are too many possible causes of anything you measure to identify a single cause and effect; overlapping marketing campaigns makes it impossible to tell what’s working and what’s not working!’’ The ‘‘answer’’ in large part is one of taking a systematic and disciplined approach to marketing campaign execution. The idea is conceptually simple: conduct a small experiment, isolating as many variables as possible, to see what works and what does not. Although the majority of marketers are aware of this approach, my research shows that the vast majority of marketing organizations, almost 70 percent, do not use experiments to pilot test marketing campaigns relative to a control group. Why? The answer is that most marketing organizations’ reward systems are based on activity, not results. Ad- dressing this cultural challenge is the focus of overcoming Obstacle 5 at the end of this chapter. For now, let me share some examples of the power of designing marketing experiments. Harrah’s Entertainment is the world’s largest casino gaming com- pany, and it routinely design experiments to quantify the impact of its marketing. As one example, it designed an experiment targeting two
34 Data-Driven Marketing groups of frequent slot machine players in Jackson, Mississippi. The experiment consisted of a control group, the standard marketing offer, and a test marketing offer, the ‘‘Challenger.’’ The control group consisted of the typical offer: a $125 package of a free room, two steak dinners, and $30 in free chips at the casino. As expected, gaming activity for this group was the same as previously. The Challenger offer consisted of $60 in free casino chips, no hotel room, and no steak dinner. Gaming activity of those who received this new offer was measurably higher than that of the control group in subsequent months. Additional experiments showed the same result in other U.S. geographies. As a result, Harrah’s was able to cut the budget for this type of marketing by more than 50 percent and increase the performance of the marketing. This type of marketing was classiﬁed as demand generation market- ing in Chapter 1. The marketing activity drives demand in a relatively short time period following the offer, often because the offer expires. Limiting who gets the experimental offer addresses the causality issue and, because it is demand generation marketing, there is a minimal time delay between executing the campaign and obtaining the results of the campaign. In the previous section, I gave the example for the Walgreens pharmacy chain and how data were used to refocus its newspaper advertising. Following this initial win, the marketing manager started to conduct experiments with different newspaper insert advertising in different zip codes. Using speciﬁc zip codes as control groups, measuring pre-post changes of sales through the channels (stores), Feldner was able to optimize the newspaper marketing. Newspaper marketing is the most traditional marketing medium, dating back hundreds of years, yet this example shows how experimental design combined with new technology (geospatial data) can be used to radically improve performance of marketing in this medium (see Figure 2.1). The use of geospatial customer data is applicable to a broad range of marketing activities. Microsoft, for example, uses similar geospatial data to isolate clusters of piracy networks and then focuses antipiracy marketing on these speciﬁc geographies. The idea is that all geographies are not equal, and marketing should focus on the highest- value areas. See Chapter 6 for other examples of value-based marketing. Another objection related to causality is that there can be a signi- ﬁcant time delay between marketing, such as branding, and actual product purchase. An executive once told me with frustration that her
Where Do You Start? 35 chief ﬁnancial ofﬁcer (CFO) wanted to see the ﬁnancial ROI of branding initiatives. Financial metrics work great for demand genera- tion marketing, which has a short time delay between action and reaction of the customer—the customer purchases as a result of the promotion or event, and the purchase can be measured with cash. But ﬁnancial metrics do not work for awareness marketing designed to build brand, which is often separated in time from the purchase. The CFO was asking for the impossible—ﬁnancial ROI simply does not work for branding.4 If your CFO is expecting a ﬁnancial ROI for all marketing activities, you may need to sit down and explain to him or her why branding and awareness is important in customer decision making and that different metrics are needed that are leading indicators of purchasing intent. The next chapter provides a framework and examples for marketing measurement linking the 10 classical metrics to major marketing activities. The takeaways are that the problem of identifying a cause and effect can be addressed by (1) using experiments to isolate out impacts and (2) using the appropriate metrics for each different type of marketing activity. Overcome Obstacle 3: Lack of Data—Strategies for Obtaining Customer Data Many marketers struggle with too much data, rather than not enough. Collecting and analyzing the right data was the subject of the over- coming Obstacle 1. However, B2B companies have a legitimate data challenge, since these companies do not sell directly to customers. They sell through a channel and do not have direct access to customer transactional data. There are three approaches to overcome this obstacle. 1. Channel Partner Data Sharing Each morning John Chambers, CEO of $39 billion network infra- structure company Cisco Systems, ﬁrst opens a can of Diet Coke and then logs on to the Cisco e-Sales portal. Through this enterprise, web application, he can drill down globally into all sales that occurred the day before, dicing by geography, company that purchased, or products, or tied to a speciﬁc sales account manager. This is a neat trick, especially
36 Data-Driven Marketing considering that Cisco sells more than 95 percent of its products indirectly through value-added resellers (VARs). How does Cisco do it? The answer is that it contractually requires VARs to share customer sales data. Most B2B companies ﬁnd that requesting customer data from the channel partner is met with a ﬁrm ‘‘no way.’’ The partner refuses on the grounds that the data are its property and the source of its competitive advantage. Cisco appears to be the exception rather than the rule, requiring channel partners contractually to share data in order to resell their products. Many B2B ﬁrms I work with tell me that the relationship (the contract) is already in place and can’t be changed. I argue that every- thing is negotiable. One ﬁrm I worked with had a large dealer network and owned 15 percent of the largest dealers: an easy solution for them was to start with the dealers they owned, show the beneﬁt of the data- driven approach, and use this as a motivation for other large dealers to share. The B2B company must answer the question: ‘‘What’s in it for my channel partner to share its data?’’ One carrot is in the fact that B2B companies often spend considerable marketing dollars comarket- ing with channel partners. Shared data analysis provides deep insights into how to radically improve the performance of this type of marketing. Microsoft, for example, found that original equipment manufacturer (OEM) partners gladly accepted and used the Microsoft-developed comarketing collateral, and in some cases were willing to share customer data to show the efﬁcacy of the marketing. Note that the B2B ﬁrm does not necessarily have to know the name and address of the customer; this could be deleted from the shared data ﬁle. What you really need is insight into what products or services the customer purchases, and the ability to act upon these data, perhaps through the channel partner. The approach of ‘‘disguising’’ the customer data can sidestep the channel partner’s concern that if the B2B ﬁrm has the customer data, it will want to go direct and cut the channel out of the deal. 2. Frequent Drinker Programs Suntory is one of the largest liquor distillers in Japan and brews a beer called Suntory Malts. The three major brands of beer in Japan are Asahi, Kirin, and Sapporo. Suntory Malts beer is in the third tier of popularity, in terms of sales revenue and brand awareness. In the late 1990s, Suntory
Where Do You Start? 37 did something with the Internet that at the time was particularly innovative: it created a frequent drinker web site. Suntory sells all of its beer indirectly through beer distributors, bars, restaurants, grocery stores, and vending machines. The idea of the web site was that consumers would come to the web site and input the number of beers they drank, given by codes on the bottles, and in return get points. With the points, customers got to purchase silly hats, bottle tops with their name on it, or the chair that is too uncomfortable to sit in. What brilliant marketing! At the peak of the campaign, Malts reported 300,000 visitors a month to the web site: these were the high- value customers, the frequent drinkers. And the web site enabled the collection of data for direct marketing of other Suntory products. I lived in Japan for two years and am intimately familiar with the vast amounts of alcohol consumption in the ofﬁce parties after work. Japan has a culture of drinking that makes this example work, whereas in the United States it is not politically correct. However, the idea of frequent drinker programs transcends cultures and geographies. Similar to Suntory, Coca-Cola has used the approach for My Coke Rewards (www.mycokerewards.com). The web site is a loyalty program for frequent Coke-branded product drinkers. Again, users get points for how many Coke drinks they have, and then get rewards such as T-shirts, DVDs, and discounts with the many partners afﬁliated with the market- ing activity. Yes, the customer data are as cheap as a T-shirt! My Coke Rewards enables Coke to have direct access to their ‘‘frequent drinkers’’ and to do direct marketing via e-mail. The portal also serves as a revenue generator for Coke through the paid ads on the site. The frequent drinker program idea is not limited to beverage manufacturers—let me give you a different B2B example. Microsoft sells almost all of its products indirectly through OEM partners and channels. As a result, except for the largest enterprise customers, it does not know who purchases its software. One market segment that is particularly important to Microsoft is the midmarket business segment. These are medium-sized companies that use Microsoft products, and Microsoft sees this segment as an area for potential future growth. The challenge, again, is that Microsoft sells indirectly to this segment and as a result does not actually know what the customers purchase. The Microsoft revenue model for midmarket and enterprise businesses is to sell annual software licenses. However, if the customer does not renew the license, uses the product, and then later goes
38 Data-Driven Marketing to purchase the upgrade, there is an additional fee for off-license product use. Microsoft created the midmarket portal, a web site where midsized customers can enter their licenses, and in return Microsoft will ensure that ﬁrms optimally manage their licenses. While the exact numbers are proprietary, Microsoft now has a signiﬁcant fraction of global midmarket companies in its database. The midmarket portal enables Microsoft to provide value to its customers—it provides a service, helping midmarket customers to manage their licenses and save money. In return, Microsoft collects data on the product purchases and can use this information for outbound marketing. As an example, Microsoft does analysis of product preferences to understand the bundles of products customers purchase and what might be a good product offer for a particular customer (see Chapter 9 for the three essential data analysis techniques). As a result of this analysis and targeted marketing, Microsoft saw a more than ﬁve times increase in performance of marketing campaigns. Note that in all of these examples there is a crystal-clear value proposition for customers or resellers to share their data. In the case of the Suntory and My Coke Rewards web sites, the value is the ability to get discounted products and services. In the Microsoft case, it was cost savings due to better software license management. For your business, ask: ‘‘What is the value proposition for customers to provide their data?’’ 3. Use Surveys as a Proxy for Customer Data The third approach is to use focus groups and surveys for ﬁne-grained segmentation and target marketing. The idea is to capture the demo- graphics, characteristics, and purchasing afﬁnities of your end customers through in-depth market research. You can then create survey-based, data-driven marketing offers targeting these segments, and again test these ideas using focus groups and experiments discussed in detail for overcoming Obstacle 2. This approach is not as effective as analyzing large customer transaction data sets, but it can be a great way to get started if you are a B2B ﬁrm and have a major roadblock of obtaining direct customer data. I discuss using surveys and focus groups to over- come the B2B data challenge in more detail in the next chapter. Professional services ﬁrms may charge $20,000 or more to design, conduct, and do data analysis for a focus group session of 10 to 15 participants. You can get started, however, for the cost of a free lunch and a gift for participants to incent them to show up. Invoke Solutions
40 Data-Driven Marketing work, and for the approach to be easy rather than painful, is to design all campaigns in advance so that they are easy to measure. It’s actually not that hard. My experience is that deﬁning the appropriate metrics and how the data will be collected usually takes only a few hours during the campaign planning stage—and at most a day or two for a very large campaign—but lays the essential groundwork for data-driven marketing during the cam- paign execution. The next chapter is a systematic discussion of how to choose the right metrics for any marketing activity and how to design campaigns for easy measurement. Designing for measurement up front is the proverbial 1 percent that yields the 99 percent of value after the campaign is done: it enables you to quantify the impact of your marketing. The great thing about data-driven marketing and marketing mea- surement is that, by deﬁnition, they create a business case for future marketing investments so that you can justify increased spending for infrastructure to support data-driven marketing activities. Infrastructure for Data-Driven Marketing The laptop computer, combined with Microsoft Excel, is an amazingly powerful tool. For infrastructure, this is all most marketers need to get started on the data-driven marketing journey. I want to manage expect- ations, though. Microsoft Excel 2003 had a limit of 65,536 rows, or customer records, in a spreadsheet, and Excel 2007 had a limit of 1,048,576 rows by 16,384 columns.5 This means that if you have a large number of customers, Excel is not going to work as a marketing database, nor should it. You want a single version of the truth, not different duplicate copies of customer data on each marketer’s desktop. Excel is great for analyzing branding and customer satisfaction survey data (Chapter 4), getting started with Internet metrics (Chapter 7), and calculating ﬁnancial return on marketing investment (ROMI; Chapter 5). You can also get started with marketing campaign score- cards in Excel (Chapter 3), and in Chapter 8 I show how Microsoft tracked a $17 million campaign using near-time weekly data in Excel. So Excel works to get started with the vast majority of metrics in this book. Where Excel does not work is as a direct marketing database (Chapters 1 and 9), for value-based marketing (Chapter 6), and for analytic marketing for a large customer base (Chapter 9). For these applications, your data requirements will drive the scale of the infrastructure you need. If your data requirements are to analyze a
Where Do You Start? 41 data set of several thousand customers, segment the customers and multiple dimensions, and then design target marketing for these cus- tomers, you are good to go with a laptop and software such as Excel or SAS JMP (see Chapter 10 for a detailed example for EarthLink with a downloadable data set and instructions). However, if your goal is to solve the same problem for 50 million customers, segment out the top 1 million using multiple dimensions, and do target marketing for these 1 million, you need industrial-strength infrastructure. Said Richard Winter, an expert in high-performance data warehouse design: ‘‘The difference in these requirements is the difference between building a ranch house or the Empire State Building.’’ Furthermore, if your plan is to do the analysis on a one-time basis, you can start using relatively low-cost systems and do a lot of the data pulling manually. But if the plan is for event-driven marketing, based on current customer purchases and the real-time calculation of their future value, then the infrastructure requirements again increase. My point is that size of the data set, which comes from the customer base and related interactions of the customer with your ﬁrm, and what you plan to do with the data, drives the requirements for the marketing infrastructure. The size of the customer base, as well as the amount of data and how often you want to analyze it, will determine how much money you need for the infrastructure. Chapter 10 discusses these infrastructure trade-offs for different customer base sizes and marketing requirements. However, I return to the essential theme of this chapter: think big, start small, and scale up fast. The takeaway is that when thinking big, there needs to be a realization that the end goal of a data-driven marketing strategy will ultimately drive the infrastructure requirements and that you need infrastructure to scale fast. Infrastructure for a Large Firm A very senior business executive in a Fortune 500 company recently exclaimed to me in a very exasperated tone, ‘‘What does it all do?’’ He was frustrated by the high cost of the data-driven marketing infrastructure technology and the inability of IT to communicate in plain English what it did. Another business manager told me, ‘‘IT couldn’t explain what they do at a party.’’ Let’s ﬁrst take a look at data-driven marketing infrastructure; I’ll deﬁne the essential components and provide a concep- tual model for how it all works. Then we can discuss how to overcome the hurdle of marketing working collaboratively with IT.
42 Data-Driven Marketing Figure 2.2 Translating data-driven marketing strategy into infrastructure for a large ﬁrm. A large ﬁrm with a large customer base requires a corresponding infrastructure, shown in Figure 2.2. Think of Figure 2.2 as the end-state infrastructure that embodies the data-driven marketing strategy. On the right-hand side are systems to collect data from the various touch points of the ﬁrm with the customer. These systems are typically called operational customer relationship management (CRM). They collect customer data from the point of sale, call centers, the web site, and customer returns. So, for example, when a customer at Jiffy Lube pulls around to the back of the oil change station, these operational CRM systems collect the license plate data and prompt the service represent- ative to say, ‘‘Hi, Mark, it’s been 4,000 miles since your last oil change,’’ just as you get out of the car. Data from each customer visit are tagged and input into the enter- prise data warehouse (EDW), shown on the bottom of Figure 2.2. In a large enterprise, the EDW is a very big data store of all customer interactions with the ﬁrm, and ideally also includes ﬁrm operational and ﬁnancial data. On the left-hand side of Figure 2.2 are the technical tools needed to mine the EDW and generate reports. Example reports might include weekly updates on district sales numbers, customers who canceled their subscriptions, and so on. In Figure 2.2, most important is the data-driven marketing on top of the EDW: analytics for segmentation, targeting, and relationship build- ing with customers. For the Jiffy Lube example, the ‘‘hello’’ at the point of sale is an example of personalized interaction marketing and would be facilitated by a program script that takes the license plate data,
Where Do You Start? 43 searches the EDW, pulls the customer record, uses a business rule to ﬁgure out what the prompt should be, and ﬁnally displays the prompt on the computer screen of the Jiffy Lube technician (see Chapter 6 for another detailed example at RBC). Analysis and modeling involves using the data-mining tools for targeted marketing. Chapter 9 gives examples of the three essential analysis techniques—propensity modeling, market basket analysis, and decision trees. For example, Meredith, publisher of Better Homes and Gardens and many other women’s magazines, uses modeling analysis to ﬁgure out which product customers are most likely to buy next. Weekly marketing campaigns are accomplished with communication and per- sonalization tools that pull the customer data from the EDW, run the offer prediction model, and then send a customized e-mail to speciﬁc customers. See Chapter 9 for the details. How do you start to build this infrastructure? Start small, with a clearly deﬁned business case, and show the how data-driven marketing enables a better, faster, cheaper, and smarter way to do something your marketing already does. Then build infrastructure incrementally with a business case for each stage. Chapter 10 answers the question ‘‘What’s it going to take?’’ for data-driven marketing infrastructure and describes the Harrah’s infrastructure story in detail. The Harrah’s example shows how it incrementally created a portfolio of value that became a source of strategic advantage in the casino gaming industry. The Marketing and IT Relationship I often ﬁnd tension between marketing and IT in companies. My per- spective is that the relationship between marketing and IT should be like going to your doctor. For example, if you hurt your elbow playing tennis, you don’t go to the doctor and say, ‘‘I want an MRI and a jar full of Vicodin to kill the pain.’’ Rather, you explain your symptoms and the doctor prescribes the solution. Similarly with IT, you have to clearly deﬁne the marketing business requirements, the objectives, what you want to do with the data, and so on. IT should then prescribe the solution and is responsible to meet your requirements—they should deliver the system in a reasonable amount of time and to the budget you agreed upon. Market- ing is responsible for the business returns of the system, not IT. For readers who have experienced a life-threatening illness of a family member—a parent, child, or other relative—a natural response is to become an ‘‘expert’’ in the illness. You seek out all the information
44 Data-Driven Marketing you can, consult experts for opinions, and become educated on the best practice course of treatment, prognosis, and associated risks. Data- driven marketing infrastructure projects are no different. As a marketer, you have to become an educated consumer of the technology that supports your data-driven marketing activities. Learn to ask the right questions to ensure that the wheels don’t fall off the project and that the system delivers value. Chapter 10 is an in-depth discussion of these issues and will give you the essential knowledge to be an educated partner in the development of data-driven marketing IT. In summary, data-driven marketing technology is too important to leave to the technologists. Overcome Obstacle 5: People and Change—Create a Data-Driven Marketing Culture I often hear, ‘‘But I am a line manager and have no inﬂuence to create change.’’ I ﬁnd that many people underestimate their potential to inﬂuence others. After all, small changes can sometimes have big impacts. In Chapter 8, I share the in-depth case example of Microsoft’s Security Guidance campaign. As we will see, just changing the landing page for the impression advertising resulted in a 400 percent improve- ment in the campaign performance. The takeaway is that small changes can have big impacts. Realize that you are a culture of one person and change starts with you. I suggest starting with changes that positively impact those around you. For example, the next chapter details how to systematically apply marketing metrics to marketing campaigns and how to create a balanced scorecard for marketing. You can apply these principles to your market- ing, train your team on how to use the scorecards, and educate your management on the value of the approach. Results talk, and the scorecards will enable you to keep score for your marketing initiatives. Of course, creating a data-driven marketing culture requires more than one person. You have to convince others, and the quick win is an essential early step. If you are down the chain of command, this may seem like a daunting task, so I asked executives what they did when they were starting out in the organization. Kelly Cook explained that when she was at Continental Airlines, she ﬁrst found out which areas of the business had the data she needed. She then made relationships with the four or ﬁve key people in the
Where Do You Start? 45 different functional areas who were like-minded. The result was that she convinced others at her level across the business to work together, and they formed an informal team that worked toward early successes. Convincing the right people requires an understanding of who has the power in your organization. Often, the most powerful people are not the senior executives. The Trouble with Change Is People If you are a small ﬁrm, given the boss’s buy-in and demonstrated results, driving change should be fairly straightforward. However, in a large organization, transforming the culture to data-driven marketing is no easy undertaking. Signiﬁcant buy-in is required, for example, to build the infrastructure of Figure 2.2, which is most likely a multimillion- dollar investment. Corporate cultures can be grouped into three major categories: rational, bureaucratic, or political. My young MBA students often believe that organizations are rational, that is, that the best ideal will prevail. Experienced managers know that this is decidedly not the case. The other two organization cultures are most likely in play: A bureau- cratic organization has a very rigid organizational structure, and protocols must be strictly followed in communicating with senior executives. These organizations are militaristic, with the general giving orders from the top, and the commanders ensuring the orders are executed on the front lines. Political organizations, in contrast, have centers of power, with individuals who have kingdoms within the organization, often accompanied by budgetary authority and staff. I work in a univer- sity and can testify that universities are the most political places on the planet. As Henry Kissinger once said, ‘‘The reason the ﬁghts are so ﬁerce in academia is because the stakes are so low.’’ Navigating a political landscape takes experience, but understand- ing who has the power in your organization is a useful ﬁrst step. Seek out senior executive sponsors who have power and appreciate results, and who like to use data-driven decision making. In a large ﬁrm to be successful, the broader initiative requires a strong senior executive guiding coalition. This executive council oversees the strategy devel- opment and the execution, and it monitors progress. There may be a tendency to make the infrastructure development (Figure 2.2) into a large IT project led by IT. This is the kiss of death for the data-driven marketing initiative—marketers, not technologists, must lead the way.
46 Data-Driven Marketing Although senior executive sponsorship is essential, buy-in from both middle and line managers is also needed for success. Why is change so hard on the front lines? People overestimate the value of what they have and underestimate the value the change will bring. The best motivator to drive change is to have a crisis6: ‘‘Our marketing budget is being cut by 36 percent; we need to justify our future marketing spending.’’ ‘‘We are losing signiﬁcant market share.’’ ‘‘Our discount marketing is killing overall proﬁtability.’’ ‘‘We are hemorrhaging customers and don’t know who are the most proﬁtable.’’ ‘‘Our competitors are consistently outmarketing us.’’ These are all motivators for the marketing organization to change. In 2009, the ﬁnancial meltdown and recession caused massive layoffs and spending cuts across the board. With a recovery of three years or more anticipated to get back to pre-recession levels, now is a great time to change to a data-driven marketing culture with the global ﬁnancial crisis as a motivator. How do you get noticed and gain a seat at the table? Everyone loves a winner. So again, start small and get the quick win. People around you will notice. Then get the next win—this builds momentum for change and inﬂuencing others, building your credibility and stature within the organization. People want to be on the winning team, so create a buzz around what you are doing and show how good things happen when you follow the principles of data-driven marketing. Creating Incentives for Change: Measurement and Behavior Like many Americans, I struggle with my body weight—I travel a lot and eat out often. The result is an expanding waistline. Yet, I have known for a long time that the equation for weight loss is simple: eat fewer calories than you burn through daily activity and exercise. But counting calories is a pain. Recently I found a free application on the iPhone called Lose It that makes counting calories easy. For each day, you can easily add foods and exercise, and copy, paste, and edit previous meals, so it takes only a few seconds. The result was data-driven transparency in my diet—I realized that my coconut chocolate chip cookie binge each evening was
Where Do You Start? 47 costing 600 calories or more. But my budget was only 1,600 calories a day to lose two pounds a week. This meant I had a decision to make: keep eating the cookies or ﬁnd an alternative. I found that half a cup of caramel praline crunch ice cream was only 160 calories—what a deal! And more satisfying to me. My point is that if you can measure something, you can control it. In this case, the measurement was the calories consumed and the exercise calories burned. The result was clarity in decision making for what I eat, which meant I lost the weight and ultimately completely changed my diet. Measurement can also change a culture, especially if you make the measurements public. Fighter pilots in the United States Navy are highly competitive individuals. There is a scoreboard in the brieﬁng room of the aircraft carrier that rates each pilot’s mission on multiple dimensions and is directly compared with scores of peers. The result is transparency in what constitutes an effective mission and peer pressure to improve individual performance. The Kellogg School of Management is a top management school. Part of the reason for this is that some 30 years ago, Dean Donald Jacobs made the student ratings of the classes ‘‘public’’ information to all students and faculty. There was an outcry from the faculty, but within several months the ratings increased very dramatically across the board: the faculty did not want to look bad compared with their peers, so the teaching quality improved. For marketing, making metrics and measurement ‘‘public’’ within the organization will incent change. But you have to measure the right things. Many organizations incent activity, not results, so the idea is to focus on the metrics that really value marketing. The next chapter is all about how to do this. If your goal is to drive new data-driven marketing approaches across a large marketing organization, given executive buy-in and air support, explain to the campaign marketers why they need to do this and demonstrate the results. Training is essential so that they have the skills to use the new approaches and tools. Ideally, there is grassroots support, and the buzz from winning marketing activities creates a snowball effect of positive change. But there will always be laggards. You need to explain the beneﬁts and approach again to those who are not doing it, and put incentives in place for success, such as bonuses for good performance. I believe in second chances, but when it comes to the third time, a big stick can be most effective.
48 Data-Driven Marketing In the mid-1990s, Harrah’s Entertainment fundamentally changed its strategy to incent customers play at multiple Harrah’s properties. This new strategy required data sharing between casino properties for data- driven marketing (see Chapter 10 for the details). The casino general managers (GMs) had been incented on the proﬁt and loss of their individual property, and some of the casino GMs were very resistant to the new strategy, since they did not want to give data to their ‘‘competi- tor’’ GMs at other Harrah’s properties. Harrah’s ultimately ﬁred a few of its top-performing GMs who were not in compliance with the new data- sharing policy—the organization instantly got the message across the board. Overcoming the Data-Driven Marketing Skills Gap My research points to a signiﬁcant data-driven marketing skills gap in organizations: 64 percent of survey respondents report that they do not have enough employees who have the skill to track and analyze complex marketing data; 55 percent of respondents said that overall their marketing staff does not have sufﬁcient working knowledge of ﬁnancial concepts such as ROI, NPV, and CLTV (Chapter 5 and 6 metrics). The personnel skills gap barrier was echoed in my interviews. For instance, one executive told me, ‘‘One of the biggest hurdles is person- nel and their ability to understand this new world of marketing. The number of people that have really deep e-marketing backgrounds plus brand backgrounds could probably measure on one or two hands.’’ Another executive said, ‘‘One of the many challenges is that there are lot of processes that still rely on human intervention and human prophecies. Whenever you have that happening, you know there are always going to be human errors.’’ The theme here is that to bring up the game of the overall marketing organization requires training. You need to give your people the new approaches, tools, techniques, and skills to optimize marketing man- agement and deliver best in class data-driven marketing. Kelly Cook told me, ‘‘You have to have a good marketing and business strategy, you also have to have the work processes behind the model and you have to have the technology tools behind that. The fourth critical component is the employees. Data-driven marketing is not a ‘‘You build it, they will come’’ model for employees. They must want to deliver exceptional marketing performance. I mean, do you know how fast I can clean my house when I want to?’’ In my experience, training is an essential component of organiza- tional change. Employee training enculturates the new techniques,
Where Do You Start? 49 approaches, and tools. Don’t skimp on the training budget—replace boring rote sessions with energizing group action learning exercises so that the participants are jazzed about data-driven marketing afterward and have follow-up coaching sessions for reinforcement. Dynamic exter- nal speakers have another beneﬁt you can use to your advantage—they can be a positive force for the data-driven marketing change initiative by credibly voicing and reinforcing your ideas to the organization. Top Down and Bottom Up The strategies given in this book enable you to have a signiﬁcant impact on your piece of the marketing organization, be it in your own day-to- day activities or on the teams you work with. My experience is that marketers who apply these principles get recognized in their respective organizations, are promoted faster, and are ultimately more successful in their careers. But cultural change for a large organization cannot be accomplished solely from the bottom up. Over the years, I have come to appreciate how equipping frontline marketers with the right tools and processes is only a piece of the equation of successful ﬁrms. For a true cultural change to a data-driven organization, senior executive leadership is essential for success. That is, the top must lead by example. The need for senior leadership sponsorship and support may seem like scaling Mount Everest in your organization. This is why understanding the political landscape of your organization is so important. You create a powerful guiding coalition by showing like-minded executives what is possible in terms of doing marketing cheaper, better, faster, and smarter by using data. You will know when you are successful when the new data- driven marketing idea is not your idea, but the senior executives. A Road Map for Implementing Data-Driven Marketing Let’s return to the question ‘‘Where do you start?’’ by deﬁning a road map for upgrading or implementing data-driven marketing (see Figure 2.3). The road map starts at Stage 1 by ‘‘designing’’ the road to the future. That is, this ﬁrst step is all about deﬁning a clear game plan. At this stage, an initial assessment of the current state is useful. Ask what metrics do you currently use. How do you use data for decision making? Given data, does your organization ‘‘‘do the right thing’’? Think through what you would like to be able to do in terms of data-driven marketing.
50 Data-Driven Marketing Figure 2.3 A road map for upgrading or implementing data-driven marketing. The road map’s next step, Stage 2, is Diagnosis, where the idea is to take the current state assessment to the next level. Ask what are the gaps and opportunities? For example, one marketing organization I worked with was focused entirely on unit sales resulting from marketing and was not measuring any forward-looking metrics. Clearly, there was a major gap, and what was needed was a balanced approach. At the Diagnosis stage, there are most likely multiple options or pathways that you could take for next steps. This is where we need to think about risk and return: of all the options, what are the easy wins that will give the highest impact at the lowest effort and cost? The short list of quick wins is the Opportunities in Stage 3 of the road map. Once the opportunities are identiﬁed, go after the easiest to get the quick win. The Stage 4 focus is on Tools: deﬁne metrics and scorecards for success and develop capabilities to support ongoing marketing activities. The quick wins are often one-time proof of concepts. The idea of Stage 4 is to put in place the infrastructure for repeatability. The last step, Stage 5, is Process—frequent reviews to evaluate performance and change course where necessary. Again you don’t need millions of dollars of infrastructure to get started. I advocate using 300 Â 500 index cards for scorecards and Excel for tracking and initial dashboards. Automate the process once you show repeatability and results. I have worked with many companies to implement the road map in Figure 2.3. Invariably, a quick 30-day assessment diagnoses problems and identiﬁes quick wins. However, implementing data-driven market- ing takes time, so focusing on the quick wins and results is a good ﬁrst step. Chapter 11 discusses the research on essential marketing processes
Where Do You Start? 51 (Stage 5 in the road map) and maturity of organizations in the data- driven marketing journey. Chapter 10 is all about the infrastructure component necessary to implement a data-driven marketing strategy, which is the Tools stage of the road map. Following the data-driven marketing road map will deliver in- credible performance improvements to your marketing, which in turn will get you noticed. Results talk, and it is much more fun to be a winner and have respect in the organization than to constantly be questioned about the value of what you do. Chapter Insights Start by collecting the right data to get the quick win—ask what is the 20 percent of data that will give 80 percent of the value? Overcome the causality obstacle by using the right metrics and experiment to test marketing ideas. Small experiments can dramatically improve marketing performance. B2B companies need a value proposition for channel partners and end customers to share their data. Don’t have the resources? Marketing measurement is the 1 percent of effort that is the 99 percent of value for justifying spending in the future. You need infrastructure to scale a data-driven marketing strategy. The quantity and frequency of customer data to analyze will deﬁne how much infrastructure you need. Data-driven marketing technology is too important to leave to the technologists. Reward results, not marketing activity—align measurement with incentives for change of the marketing organization and train employees to use the new tools and approaches. Senior executive leadership is needed for cultural change across a large marketing organization. There is a road map for upgrading or implementing data-driven marketing: start with assessment, quick wins, then the develop- ment of tools for repeatability, and ﬁnally add a ﬂexible review process to act on the results.
CHAPTER 3 The 10 Classical Marketing Metrics W hen I ﬁrst started teaching marketing measurement, I was told, rather ﬁrmly, by the executive participants that I ‘‘did not understand.’’ I had to agree—I was new to the whole marketing thing, since my background was in technology and data analysis. ‘‘Please explain,’’ I said. ‘‘Marketing is creative,’’ I was told, ‘‘and you can’t measure creativity.’’ My perspective was, and still is, that you can measure everything. Measurement is powerful, and as we saw in the last chapter, the act of measurement can not only radically improve marketing performance but also dramatically change the behavior of your organization, provided you measure the right metrics in the right way. So I asked in my survey research, ‘‘Do you outsource the creative component of your marketing?’’ The answer was that 72 percent of ﬁrms surveyed outsource the creative. This result has signiﬁcant implications. The vast majority of marketing organizations are not in the creative content business, but instead manage the process of marketing. Optimizing marketing processes, and the four marketing processes to focus on, is the subject of Chapter 11. 52
The 10 Classical Marketing Metrics 53 I am well aware that most organizations and marketers struggle to measure their marketing activities, given the hundreds of possible metrics. As an example, I once engaged with a major Fortune 100 company, and it shared its scorecard that contained more than 50 metrics. This scorecard took considerable time and effort to put together each month and was providing little beneﬁt; there was too much data, and these data did not provide managers with information they needed to make decisions. Clearly, what is needed is a simple approach to think through which metrics are important for a speciﬁc type of marketing activity. Linking Marketing Activities to Metrics The fact that an idea is old does not mean that it is not good. The marketing behavioral impact model, sometimes called the purchasing funnel, was ﬁrst published in the 1960s. The idea is that different marketing activities take the customer through the stages of awareness, evaluation, trial, and loyalty. That is, marketing activities are designed to ‘‘funnel’’ customers from awareness to ultimately become loyal customers. This 40-plus-year-old idea has new signiﬁcance today as technology enables measurement across this spectrum like never before. Figure 3.1 is my modern interpretation of the marketing behavioral impact model. In this picture, the purchasing ‘‘funnel’’ is a continuous cycle where loyalty feeds awareness. Let’s review this cycle from a modern marketing measurement perspective and also make the con- nection to the ﬁrst 10 essential marketing metrics—the ‘‘classical’’ metrics. Awareness Marketing Awareness marketing comes in many forms such as TV advertising, billboards, sports sponsorship, naming rights to stadiums, print adver- tising emphasizing a brand, and creative use of the Internet. Awareness and branding are intimately related. Simply put, a brand is a consumer perception of a particular product or service and may encompass the whole company, such as Disney or Apple. This perception is driven both by marketing and experience with the product and by recommendations of friends and colleagues. Branding is incredibly important, because it often drives the consumer to take a ﬁrst look at your product or service
54 Data-Driven Marketing Figure 3.1 The marketing behavioral impact model with example marketing activities. and can have the advantage of enabling a ﬁrm to charge a price premium over nonbranded competitors. In the purchasing cycle, awareness is the furthest removed from the customer purchase, and there can be a signiﬁcant time delay between awareness marketing and actual sales. Hence, ﬁnancial metrics are not particularly useful for measuring awareness and brand marketing. Firms often conduct large brand awareness surveys, which track customer awareness across geographies and over time. These qualitative data are collected using large sample surveys, 350 people or more in each segment and geography, and are very costly and time consuming. Hence, large organizations undertake their brand survey once or at most twice a year. In addition to brand awareness surveys, typical metrics to measure efﬁcacy of awareness marketing are number of attendees at events, eyeballs on a web site, or media impressions. Joyce Julius, for example, measures sports sponsorship brand exposure. They have sophisticated systems that track where the corporate branding from the sponsorship appears on TV, and they then calculate the equivalent cost of purchas- ing this TV advertising time.
The 10 Classical Marketing Metrics 55 Tiger Woods’s win in the 2005 Masters, for example, received $10.4 million of TV exposure for the Nike logo, and Jeff Gordon received $9.9 million of exposure for DuPont brands when he won the 2005 Daytona 500. In total in 2005, DuPont received $85 million in TV exposure due to the Jeff Gordon sponsorship. The challenge with these metrics is that they do not connect to purchase intent and do not capture efﬁcacy of the marketing. The chief marketing ofﬁcer of DuPont, David Bills, voiced his frustration to me with this measure- ment approach for DuPont’s sponsorship of Jeff Gordon in NASCAR: ‘‘DuPont is fundamentally a business-to-business company and we would not ordinarily spend $85 million on consumer advertising.’’ Clearly, there is a gap between impressions and marketing value. So what is the essential metric? The key metric is ability of the customer to recall your product or service: Metric #1: The Essential Awareness Metric Brand awareness ¼ Ability to recall a product or service Top-of-mind recall means that in the purchasing cycle (Figure 3.1), your product or service will be one of the ﬁrst the consumer thinks about to consider purchasing. There are a few more sophisticated metrics related to metric #1, but all are essentially a measure of the ability of a customer to name a company or product. I discuss branding impact and how to actually measure awareness in more detail in the next chapter. But what if you do not have the resources of a large ﬁrm or the time to wait for the global brand awareness survey results? The new medium of the Internet and/or cell phone text messaging can be used to connect awareness, trial, and demand generation marketing. In a sports stadium, for example, placing a uniform resource locator (URL) or text message number on a billboard can quantify the impact of the marketing. I advocate that all TV, print, and billboard advertising should have a URL or texting number. By slightly changing the URL or text message number, one can quantify how many people acted as a result of the awareness advertising. In Chapter 8, we will come back to this concept in detail and show how you can use these techniques to build agility into the design and execution of a marketing campaign. This agility can improve perform- ance by factors of ﬁve or more.
56 Data-Driven Marketing Evaluation Marketing Evaluation marketing is designed to drive customer purchase intent by enabling customers to compare different products or services. Examples include product white papers, print ads with a breakdown of the beneﬁts and features, product brochures, and web sites with product descriptions. As a speciﬁc example, Dell predominantly competes based on price because it has a low-cost direct channel to the customer (its web site) and outstanding supply-chain management that drives low-cost man- ufacturing. The price of the product is prominently featured in Dell’s evaluative marketing, and it tends to take a commodity approach to both the product and advertising, providing a list of the facts. This is a good approach if the consumer’s primary criterion is price, allowing a quick evaluation of competing product features and the customer to weigh the price performance trade-off. Apple takes a different approach in its evaluative marketing, emphasizing the cool designs and product innovations. Apple iPhone ads highlight the innovative beneﬁts of the technology, such as the App Store that has many thousands of apps for every conceivable need. Speciﬁcally for laptops, Apple charges a price premium over Dell, and tends to deemphasize price in the evaluative marketing: you don’t see the price on the Apple web site, for example, until you actually start to custom conﬁgure your computer. Evaluative marketing articulates the value proposition of the prod- uct or service, the beneﬁts, and cost trade-offs. There are many ways to present the relevant information to the customer, but there are com- monalities in marketing measurement. A challenge for evaluative marketing measurement is that there is a time delay between evaluation and purchase that could be weeks, months, or longer, depending on the product. Another related challenge is linking the evaluative marketing to an actual purchase. For these reasons, ﬁnancial metrics do not work particularly well for evaluative marketing unless one can track who looked at the evaluative marketing and then subsequently purchased. Standard metrics for evaluative marketing include product information downloads for a web site or impressions of evaluative marketing print ads. But these metrics do not measure the impact of the evaluative marketing particularly well. So how does one quantify the efﬁcacy of evaluative marketing? The answer is to ﬁnd metrics that point to future sales. Anyone who has purchased a car, even a used car, has probably gone to a dealer and picked up brochures for new cars that ﬁt their purchasing
The 10 Classical Marketing Metrics 57 criteria, and then done a side-by-side comparison of these glossy picture- ﬁlled documents. These brochures, and the related web sites, are examples of evaluative marketing in the auto business. What is the value of the glossy brochure? That is hard to quantify; however, we can deﬁne a metric that is a measure of future purchasing intent that embodies the collective impact of the evaluative marketing activities: the test-drive. It turns out that someone who test-drives a car is very likely to purchase the car. The purchase probability is not 100 percent given a test- drive; instead, there is a probability of purchase. By measuring the number of test-drives and the number of customers who subsequently purchase, one can calculate this average probability of purchase; it’s just the number of purchases divided by the number of test-drives. Another related metric worth measuring is foot trafﬁc in a car showroom, since higher foot trafﬁc should result in more test-drives, of which a fraction will convert to sales. The idea is similar to how American football coaches keep score. Their key metric is not the actual score on the scoreboard but the number of ﬁrst downs (each 10 yards the team advances on the ﬁeld). Given enough ﬁrst downs, the team should get a high score and win the game. Test-drives are a leading metric that point to future sales. In the auto business, evaluative marketing activities should therefore be designed to increase the number of car test-drives and showroom foot trafﬁc. One can therefore design experiments and measure foot trafﬁc and test-drives resulting from speciﬁc evaluative marketing activities and optimize based on this metric. One can also utilize focus groups and qualitative methods to estimate ‘‘intent to purchase’’ following exposure to different eval- uative marketing materials, such as the new car brochures. I discuss the test-drive metric in more detail in the next chapter. As we will see, the test-drive is applicable to much more than just cars, and I will give several examples, including Intel chip sales, sunglass purchases, and medical system sales. Metric #2: The Essential Evaluative Metric Test-drive ¼ Customer pretest of a product or service prior to purchase Loyalty Marketing Loyalty marketing activities may include customer assistants such as the Nordstrom Concierge Service to high-value customers. Other examples
58 Data-Driven Marketing include proactive event-driven marketing, such as Jiffy Lube sending customers a marketing offer for an oil change soon after the customer has driven 3,000 miles. In addition to repeat sales, a key metric for loyalty is churn rate: Metric #3: The Essential Loyalty Metric Churn ¼ Percentage of existing customers who stop purchasing your products or services, often measured in a year Customer churn is a particularly interesting metric and impacts some industries more than others. The U.S. cell phone industry, for example, has a churn rate on average of 22 percent per year. I once shared this statistic with an executive from a major South American telecommunications company, and he exclaimed, ‘‘Wow, that’s really good!’’ I inquired what churn he experienced in South America. The answer: 50 percent per year. I have difﬁculty imagining the challenge of potentially losing all the ﬁrm’s customers in two years. For loyalty marketing, there may be a signiﬁcant time delay between the marketing activity and the repurchase, especially if the product is something like a car, a computer, or a washing machine—a product with a long product life. This is in part why churn is such an important metric; reducing the average annual churn rate over the product life cycle can be directly translated to improving annual sales, although there may be a time delay before the impact is realized. Firms that do not know who their customers are often do not know what their churn rate is. As a result, once these data are obtained, the measured churn can come as a big surprise. For example, I worked with one major company that did not think it had a churn problem, only to ﬁnd that some parts of the business had customer churn as high as 45 percent. As we will see in the next chapter and in Chapter 6, marketing activities for retention of high-value customers can have a very signiﬁ- cant impact on ﬁrm proﬁtability. The Golden Marketing Metric: Customer Satisfaction What is missing from our discussion of awareness marketing measure- ment is a metric tied to awareness, which is a leading indicator of future sales. This essential metric is customer satisfaction (CSAT). CSAT is
The 10 Classical Marketing Metrics 59 Figure 3.2 Metrics for measurement at each phase of the marketing cycle. not the same as awareness and is more closely related to loyalty. But loyalty and awareness are intimately related; as seen in Figure 3.2, loyalty feeds awareness in the purchasing cycle. Indeed, large ﬁrms have established customer bases, and these customers have experiences with the product or service that deﬁnes their perception of the brand. As an example, a major automaker measured CSAT and brand purchase intent and found a one-to-one correspondence between CSAT and repurchase intention. Interestingly, consumers who had problems with their car were more satisﬁed and had higher repurchase intent than those that did not have a problem. Why? The excellent customer service if there was a problem with the new car measurably changed perception positively toward the brand. CSAT is therefore the ‘‘golden’’ marketing metric that bridges both loyalty and brand awareness, and it can be used as a leading indicator of future sales. Measurement of CSAT is best done by asking customers a simple question: Would you recommend this product or service to a friend or colleague? On a 10-point scale, only those who circle a 9 or 10, for deﬁnitely recommend, are highly satisﬁed and loyal customers.
60 Data-Driven Marketing Metric #4: The Golden Marketing Metric CSAT ¼ Customer satisfaction measured by asking, ‘‘Would you recommend this product or service to a friend or colleague?’’ There are a few derivative CSAT metrics, such as Net Promoter Score, but the essential measure is given above. The next chapter describes additional examples of CSAT measurement in practice. The Essential Marketing Operations Metric At this point, it is useful to introduce a metric that quantiﬁes the performance of marketing campaigns. There are several metrics that can be deﬁned to measure operational performance of a marketing cam- paign, such as cost; spend managed per employee; on-time, on-budget delivery; and so on. Many are important and can be tracked, but from an essential metrics perspective, I focus on one: Metric #5: The Essential Operational Effectiveness Metric Take rate ¼ Percentage of customers accepting a marketing offer For example, if you have 100 marketing offers sent by direct mail, telemarketing calls, TV ads, and the like, and 3 people out of the 100 who receive it accept the offer, then the take rate is 3 percent. Take rate deﬁnes how well the marketing is working from a tactical perspective, and focusing on increasing the take rate can dramatically improve marketing performance. Take rate is most often applied to demand generation marketing, discussed in the next section. However, take rate is applicable to any marketing that has a call to action, that is, a clear activity that the marketing is intended to produce from the customer. For example, the call to action for an evaluative campaign might be for customers to download a 10-day free trial of a software product. The take rate to this call to action can be measured from the number of marketing impres- sions delivered and the number of downloads. By the way, in Internet marketing, click-through rate (CTR) multiplied by transaction conver- sion rate (TCR) is the take rate measured by clicks on Internet impressions. We will discuss Internet metrics in detail in Chapter 7.
The 10 Classical Marketing Metrics 61 Demand Generation (Trial) Marketing Trial marketing in Figure 3.2 is synonymous with demand generation marketing, discussed in Chapter 1. These are marketing campaigns that drive sales in a relatively short time period. Examples include grocery store coupons that expire in 30 days, a limited-time sale for 10 percent off, or when GM offered all customers the GM employee discount. These types of marketing activities drive revenues, drive unit volume, and result in sales. Another key metric at the trial stage is lead conversion, which also results in revenue. Since all public companies are required to report their sales and net income quarterly, this type of marketing is the easiest to measure and can be quantiﬁed using cold, hard cash. That is, demand generation marketing is quantiﬁed using ﬁnancial return on marketing investment (ROMI). Chapter 5 is a deep dive into ﬁnancial ROMI and gives both an introduction to ﬁnancial concepts and detailed examples for marketing with Excel templates. For now, let’s just list the four essential ﬁnancial metrics. Taken together, these metrics enable you to quantify both demand generation (trial) marketing and new product launch marketing. The four essential ﬁnancial metrics: #6: Proﬁt ¼ Revenue À Cost #7: NPV ¼ Net present value #8: IRR ¼ Internal rate of return #9: Payback ¼ The time for a marketing investment to pay back the cost of the initiative Note from Chapter 1 that when we examined the portfolio of marketing spending, on average approximately 50 percent of marketing budgets go to demand generation (trial) marketing. In many respects, the end action of loyalty marketing is similar to trial marketing; the customer repurchases a product or service, which results in cash. As a result, loyalty marketing is also often quantiﬁable using ﬁnancial metrics. The challenge, however, is that one has to know who purchased the product in the ﬁrst place. If you don’t know who previously purchased your product, then all loyal repurchases look like trials. The trick is to know your customers; this can be particularly challenging for business-to-business (B2B) ﬁrms, and I discussed solutions to this challenge in the previous chapter.
62 Data-Driven Marketing The takeaway is that since demand generation, new product launch, and loyalty marketing drive measurable sales revenues, you can use ﬁnancial ROMI more than 50 percent of the time for marketing. This is a signiﬁcant insight: ﬁnancial ROMI is applicable to the majority of marketing activities. Of course, ﬁnancial ROMI is not ‘‘the answer’’ for all marketing measurement, and I advocate taking a balanced approach with multiple metrics in the next section. The point is that ﬁnancial metrics are applicable to marketing more often than not. The framework in Figure 3.2 provides a simple and useful guide to ﬁgure out what marketing metrics to use depending on the type of marketing activities. For simplicity, so far in this chapter we have discussed the ﬁrst nine essential metrics; the other essential metrics can be overlaid on the framework in Figure 3.2. Most important is the idea that the type of marketing activity drives what type of metrics to use. Often, marketing campaigns have multiple objectives and may incor- porate two or more elements of Figure 3.2, so for each individual campaign a scorecard of multiple metrics deﬁnes the value across the spectrum of marketing activities. A Balanced Scorecard for Marketing As you drive along a road, there are multiple sensory inputs. You look through the windshield to see hazards ahead. The dashboard with the speed and tachometer are metrics that complement what you see and help you determine whether you are driving too fast or too slowly. The rearview mirror provides sensory feedback on what is behind; the mirror gives backward-looking input. The temperature gauge and oil pressure gauge are additional operational metrics that measure how well the engine is running, and the fuel gauge provides information so you don’t run out of gas. In marketing, measuring only sales revenue is like driving a car by looking only in the rearview mirror, because sales measures what has happened in the past. What is needed is a balanced set of metrics, or scorecard, similar to the complete set of sensory inputs when driving a car. The balanced scorecard was ﬁrst made famous by Kaplan and Norton,1 and their original work deﬁned four categories of metrics for the ﬁrm: ﬁnancial, customer, internal processes, and growth and innovation. For marketing, we can use a similar approach where the
The 10 Classical Marketing Metrics 63 marketing measurement framework of the previous section forms the foundation. For marketing, an example of an essential metric that looks into the future is customer lifetime value (CLTV). This metric quantiﬁes the future proﬁtability of a customer and rounds out the 10 classical metrics for marketing: Metric #10: The Essential Customer Value Metric CLTV ¼ Future value of a customer Chapter 6 is devoted entirely to this metric and value-based marketing decision making. Note that CLTV is a forward-looking metric, since it quantiﬁes the future value of a customer. Continental Airlines, Royal Bank of Canada, Harrah’s Entertainment, and many other ﬁrms given as examples in this book make extensive use of CLTV in real-time marketing decisions. The scorecard for a marketing program or campaign will be speciﬁc to the type of campaign. However, the three broad categories of metrics for a marketing balanced scorecard are strategic (leading/forward), tactical (backward/lagging), and operational (internal). Strategic met- rics are forward-looking metrics and include branding, awareness, and customer satisfaction. They also includes test-drive metrics for eval- uative marketing initiatives and CLTV for predictive modeling of customer future value. Financial metrics are applicable to demand generation marketing and some customer campaigns. Operational metrics are internal to the marketing function and gauge how well the campaign is running from an operational perspective. The actual metrics for a speciﬁc marketing initiative scorecard will depend on the type of marketing campaign and the context of the business. Figure 3.3 is a summary of the marketing balanced scorecard with some example metrics. In order to deﬁne a scorecard that works, marketers ﬁrst should think through the marketing campaign objectives and understand where their marketing activities ﬁt in the behavioral impact model in Figure 3.3. The campaign activities follow from the big-picture plan, and metrics are then tied to the activities. For example, if a campaign has dual goals to drive awareness and sales revenues, the metrics focus should be on brand awareness and CSAT, the key ﬁnancial metrics for demand generation
64 Data-Driven Marketing Figure 3.3 A balanced scorecard for marketing. marketing to drive sales, and take rate for operational performance, along with campaign-speciﬁc cost and performance measures. Figure 3.4 shows the big picture of how marketing scorecards for speciﬁc campaigns and marketing activities are connected to an overall chief marketing ofﬁcer (CMO) balanced scorecard. The idea is to deﬁne common performance metrics that roll up to the executive level. At the campaign level, the scorecards contain both common and campaign- speciﬁc metrics. As an example, in the early 1990s MasterCard faced signiﬁcant competitive pressure from Visa in global credit card business and, as a Figure 3.4 Marketing organization scorecards.
The 10 Classical Marketing Metrics 65 result, chose to sponsor FIFA, World Cup soccer.2 The strategic vision for the sponsorship marketing was clearly deﬁned: ‘‘To underscore Master- Card’s transformation from a U.S.-oriented credit card company to a truly global brand and payment services organization,’’ said McKeveny, VP Global Promotions, MasterCard International. MasterCard deﬁned the following business objectives tied to its sponsorship strategy: Build brand awareness—through the events’ television reach and brand exposure of 7.5 minutes per 90-minute broadcast, an adver- tising cost estimated at $0.40 per 1,000 viewers reached. Stimulate card usage and acquisition—by exploiting the global appeal of the World Cup. Members worldwide were given the opportunity to implement customized marketing programs targeted at speciﬁc usage, activation, and acquisition objectives. Provide business opportunities for members—through in-branch programs to build trafﬁc, cross-sell other member products such as Maestro and MasterCard Travelers Checks, increase automated teller machine (ATM) usage, and execute merchant-driven pro- motions to increase acceptance and preference for MasterCard products. Enhance the perception of MasterCard as a global brand and payment system—by associating for the long-term MasterCard and the world’s leading sport. It is interesting to note that MasterCard is actually a B2B company since its card is licensed to member banks that act as MasterCard’s channel partners. The business objectives of the marketing program not only included end consumers, but also speciﬁcally included the member banks. Also note that while this was fundamentally an awareness and branding marketing program, an additional objective was to drive sales revenues through card acquisition and usage; this is demand generation marketing measurable by ﬁnancial metrics. The business objectives drove a tactical execution plan that evolved over several years of the FIFA World Cup sponsorship, and resulted in a scorecard measuring pre- post change across geographies. The scorecard had two different sets of metrics for consumers and member banks. For consumers, these metrics focused on brand and aware- ness and pre-post surveys asked questions to measure the brand perception change: Did the sponsorship close the awareness gap compared with Visa?
66 Data-Driven Marketing Did sponsorship awareness increase? And was sponsorship awareness higher than or equal to that of Visa during the Visa Olympics? For all of these measures, MasterCard had signiﬁcant and measurable pre-post changes across the vast majority of geographies. The scorecard for member banks was different and focused on how well the sponsorship built business-building opportunities and drove card acquisition and revenues. That is, the member bank scorecard focused primarily on demand generation metrics. Pre-post member surveys across geographies were complemented by card acquisition and revenue data. The results of the program were impressive. Brand awareness measurably increased globally, and card acquisition and usage very signiﬁcantly increased, driving measurable revenue and proﬁtability for MasterCard. Member banks also contributed to the campaign. More than 450 MasterCard members, including 75 percent of the top 100 card issuers, participated in one or more aspects of the global promotion. Together they invested $38 million in sponsorship-related marketing. Member satisfaction was measured as part of the scorecard, and 87 percent of the members stated that the sponsorship added value to their own marketing programs. This example illustrates conceptually where the scorecard ﬁts in the marketing program design and execution. Best-practice marketing program design begins with deﬁning the overarching strategic vision for the campaign. This vision is then translated into key business objectives (KBOs), and these objectives lead naturally to tactical execution. The balanced scorecard provides metrics tied to the key business objectives and tactical execution. All campaigns should have a clearly deﬁned vision statement and KBOs supporting this strategic vision. Given the tactical execution plan for the campaign, the score- card measures pre-post change on key metrics that deﬁne the KBOs. A useful exercise is to take the scorecard challenge. Early in a campaign design, have the marketing team spend a few hours brain- storming the correct metrics for the scorecard and think through how they will measure these metrics. Note that you want to focus on creating a scorecard with the few metrics that point to value—think 4 or 5 metrics, and not more than 10, for a speciﬁc campaign. I have conducted this exercise many times with executives, and invariably teams come back with a well-thought-out scorecard and measurement methodology focusing on the 3 to 5 metrics in each category that deﬁne real value.
The 10 Classical Marketing Metrics 67 If marketing campaigns are designed for measurement up front, the impact is easily captured downstream. However, if the vision, KBOs, metrics, and criteria for success are ill deﬁned at the outset, the resulting value of the campaign will be nebulous. So creating the scorecard is the proverbial 1 percent of effort that gives the 99 percent of value later. If the campaign is successful, the scorecard provides factual data to support follow-on funding. If things do not go well, however, the scorecard can provide advanced warning that you are driving a marketing campaign off a cliff. In summary, creating a marketing campaign scorecard is pretty easy. Given a couple of hours of thought, I have no doubt you can come up with one for your speciﬁc campaigns—you can use Figure 3.3 as a starting point. Yet the ability to measure and keep score is an essential ﬁrst step in the data-driven marketing journey. Marketers who system- atically measure are promoted faster and are sought after when times are tough, since they can show the value of what they do and have a track record documented with scorecards. Facing the B2B Measurement Challenge B2B companies are one more step removed from a customer. They often ‘‘sell’’ to original equipment manufacturers (OEMs), who then sell the integrated products through a value-added reseller (VAR) or other channel (possibly retail or Internet) to the end customer. Most often, B2B ﬁrms that sell through OEMs do not know the end customer, and this is often cited as a major reason why experimental design and marketing measurement are difﬁcult for B2B ﬁrms. Marketing measurement for B2B is challenging, but through an example I will show that it can be done. Microsoft, for example, sells the vast majority of its products indirectly. Its software is loaded onto personal computers manufactured by OEMs such as Dell, HP, Sony, and others, and then sold by these OEMs through channels such as the Internet, Best Buy, Wal-Mart, and other retail stores. A major challenge for Microsoft (as mentioned previously) is that it does not know who actually buys its products. Furthermore, Microsoft spends a signiﬁcant amount on OEM partner comarketing. OEM ‘‘evaluative’’ marketing is designed to help consumers eval- uate which is the best product from a number of possible choices.
68 Data-Driven Marketing Figure 3.5 Experiential test print advertisement for the Windows 7 new product launch. Source: Microsoft Marketing. Typical OEM evaluative ads often have a dense list of the features and functionality of the PC. The OEM comarketing is embedded in the ad through ‘‘[the OEM] Recommends Microsoft [Product],’’ but is this marketing delivering the best bang for the buck for Microsoft? To ﬁnd out, Microsoft regularly conducts experiments comparing evaluative advertising with more experience-based advertising. See, for example, the ‘‘experiential’’ test advertisement in Figure 3.5 for the Windows 7 new product launch in 2009. (ConToSo is a ﬁctitious OEM name for experiments.) The new product launch marketing focuses on
The 10 Classical Marketing Metrics 69 Figure 3.6 Data from the experiential versus traditional Microsoft OEM Media Center Edition print advertising experiment. Source: Microsoft Marketing. three major messages: Windows 7 (1) simpliﬁes everyday tasks, (2) works the way you want, and (3) makes new things possible. The experiential ad in Figure 3.5 highlights a few essential elements of these three messages. That is, Windows 7 is faster and more reliable, and has expanded wireless capability, enhanced multimedia, and good energy efﬁciency, which, taken together, will change your experience and make you ‘‘view your PC in a whole new light.’’ Figure 3.6 shows actual data for the Microsoft Media Center Edition (MCE) evaluative versus experiential advertising experiment. The MCE software is designed to help users manage all of their multimedia on a PC, making you ‘‘the center of your media universe.’’ For this example, Microsoft conducted an experiment consisting of four differ- ent cells. In each cell, 350 people were individually exposed to different print marketing and asked pre-post questions. Participants in the control group were exposed to traditional feature-based ads. Participants in the three other cells were exposed to ‘‘experiential ads’’ such as that shown in Figure 3.5. In these three cells the branding was varied on the ads so that participants in the three experimental cells were exposed to (1) OEM-only branding, (2) Microsoft-only branding, or (3) OEM and Microsoft joint cobranding advertising, respectively. The results are shown in Figure 3.6. For conﬁdentiality, the actual numerical changes are not shown. Instead, interpret ‘‘À’’ as a small average negative pre-post percentage change, ‘‘þþ’’ a small to moderate
70 Data-Driven Marketing positive response, ‘‘þþþ’’ a very good pre-post percentage change, and ‘‘þþþþ’’ a very signiﬁcant positive change—where pre-post is meas- ured by comparing questions asked before and after viewing the print ad. These questions include ‘‘What is your intent to purchase a Windows PC?’’ and ‘‘What is your intent to use Microsoft MCE on your next PC?’’ The results are very interesting. First, you see that the control group, the top line of data, faired okay but not great. The challenge, of course, is that feature-based ads are hard for the customer to decipher; they focus on feeds, speed, bits and bytes, and do not clearly deﬁne the beneﬁts of the MCE. The experiential ads, in contrast, articulate what the software actually does; it helps you control all of the media in your life. In the pre-post change data, the experiential ads clearly fare a lot better than the feature-based control group. But what makes this experiment particularly noteworthy is the investigation of which branding is optimal. That is, is it better to just have an OEM-branded ad, to have a Microsoft-only ad, or to cobrand with the OEM partner? The cobranded OEM and Microsoft experiential ads clearly have the best results, with likelihood to purchase and intent to use rating at the highest end of the scale, and post-message questions for brand and logo recognition rating very high also. This is an example of how 1 plus 1 sometimes equals 10. The OEM- and Microsoft-only branded ads rated poorly on the dimensions of ‘‘View Win XP Favorably’’ and ‘‘Intent to use XP.’’ However, consumer perception was signiﬁcantly more positive when the only difference in the ad was the cobranding of the logos. This example illustrates multiple important points about marketing measurement and experimental design. First, for a B2B ﬁrm, customer surveys can be used as a proxy for actual customer data. Second, nonﬁnancial metrics can also be used to point to future value. We will discuss this idea further in the next chapter, but for now note that ‘‘intent to purchase’’ serves as a pointer to actual purchasing intent and to the efﬁcacy of this evaluative marketing. As a result of these experiments, Microsoft committed signiﬁcantly more funding to OEM comarketing activities, but ensured that the funding was used to optimize the marketing impact as deﬁned by these experiments. The reaction of the OEM partner was extremely positive, as Microsoft added value in the relationship by showing how to optimize the comarketing impact and increase proﬁtability for not just Microsoft but the OEM partner as well. I have to admit that when I ﬁrst started on the marketing measure- ment journey, I had great skepticism about using qualitative data, such
The 10 Classical Marketing Metrics 71 as ‘‘intent to purchase,’’ from surveys and focus groups. This was due to my heavy quantitative background, coming from science and engineer- ing, where I was used to calculating exactly the ‘‘right’’ answer. But as I conducted more survey-based research, I realized that in marketing it is better to be approximately right than exactly wrong, and I started to appreciate the value of qualitative data. For sure, if you ask one or two people their ‘‘intent to purchase,’’ their opinions are not reliable. However, if you ask 350 people the same question, after exposing them individually to the same advertising, their responses have statistical signiﬁcance. So the takeaway is that qualita- tive data can be extremely useful, but you have to be careful that the sample size is large enough. A rule of thumb for survey-based research is that having more than 100 people in a sample is pretty good, and 300 or more is very good. However, individual in-depth interviews, small intimate focus groups of 6 to 10 people, and quick online surveys with responses from 30 to 50 customers can be invaluable in gaining customer insights. So small samples can be used effectively to rapidly test and experiment with marketing, with the understanding that the insights may be limited, and/or not generalizable to much larger samples or different geographies. I believe that it is inﬁnitely better to fail fast via experiments, rapidly adjusting to ﬁnd what works, than to crash and burn millions of campaign dollars without any measurement. Chapter Insights A spectrum of metrics exists for marketing, depending on the type of marketing activity. Financial metrics quantify more than 50 percent of marketing activities. When ﬁnancial metrics do not work, deﬁne marketing metrics that are leading indicators of sales. Take the scorecard challenge and design all marketing programs and campaigns to be measured using a balanced scorecard approach. B2B companies can use surveys and focus groups effectively to capture metrics in place of direct customer data.
PART II 15 Metrics to Radically Improve Marketing Performance
CHAPTER 4 The Five Essential Nonﬁnancial Metrics #1—Brand Awareness, #2—Test-Drive, #3—Churn, #4—Customer Satisfaction (CSAT), and #5—Take Rate Shaping Perception: Metric #1—Brand Awareness Branding is one of the most fascinating and unique aspects of marketing, because branding is all about customer perception. As an example, let’s consider bottled water. Pure water is an odorless, tasteless liquid made from molecules of two hydrogen atoms and one oxygen atom (H2O), and approximately 70 percent of the Earth’s surface is covered in water. My point is that, as products go, water is pretty much on the low end of complexity, and it’s fairly abundant. Yet, the proliferation of brands of 75
76 Data-Driven Marketing bottled water is amazing: Ice Mountain, Aquaﬁna, Geyser Peak, Poland Spring, and Dasani are a few that are top of mind for me. So why spend $2 a bottle for the brand and not 25 cents for the generic grocery store brand when the products are identical? I say this and my students have a violent reaction: ‘‘It really is different—my water comes from a mountain spring high in the Alps,’’ they say. Yep, and I’m sure the million bottles a month were ﬁlled by a guy hand- dipping them in a stream. I agree that Perrier is different—it has bubbles and it has a cool green glass bottle. Dasani claims to have ‘‘minerals’’ as a differentiator, and they went with a blue plastic bottle. Bottled water emphasizes the power of brand—the product conveys a feeling, an experience, and perception of repeatable quality that the customer is willing to pay a premium for. The value of a brand, or brand equity, is particularly difﬁcult to quantify in hard dollars. Brand equity is sometimes estimated by subtracting all tangible assets of a ﬁrm, derived from various approximations, from the market valuation of the ﬁrm based on stock price—the remaining intangible value is attributed to the ‘‘brand.’’ The challenge with this approach is that there are so many unknowns that end up multiplying together, that the derived ‘‘brand equity’’ is almost a random number. A better approach is to use surveys to ask how much more a person will be willing to pay for a branded versus the equivalent nonbranded product. This percentage difference multiplied by the product sales is approximately the value of the brand. The reader should be cautioned, however, that this approach is just an estimate, and in my opinion, the value of marketing to enhance the brand is practically impossible to accurately measure using ﬁnancial metrics. I, therefore, take a different approach, focusing on the essential nonﬁnancial metric to measure brand marketing effectiveness, and I show how to use this metric to optimize brand-marketing campaign performance. Consumer Product Brand Marketing From a purchasing perspective, a strong brand awareness gives your product or service ﬁrst consideration, or ﬁrst look, as the customer moves into the evaluative stage of the purchasing cycle (see Figure 3.2). Over time, customer perception about brands changes, and hence branding is an ongoing activity. For example, Philips Electronics manufactures electric shavers that produce an extremely close shave. Over time, electric shavers have waned in popularity, and there is a perception that
What is data-driven marketing in digital marketing? ›
Data-driven marketing is the approach of optimising brand communications based on customer information. Data-driven marketers use customer data to predict their needs, desires and future behaviours. Such insight helps develop personalised marketing strategies for the highest possible return on investment (ROI).How effective is data-driven marketing? ›
Market research firm Invesp reports that marketing firms that exceed their revenue goals apply personalization methods 83% of the time. In addition, businesses that use data-driven personalization recorded between five and eight times the return on investment (ROI) on their marketing budgets.Why do we need data-driven marketing? ›
Data-driven marketing helps businesses deal with new information and improve the quality of their products and services to keep up with changes in the marketing environment. Being data-driven also improves the overall quality of your content, so long as you're refreshing your database often enough.How do I start data-driven marketing? ›
- Gather and Centralize Data About Current Customers.
- Analyze the Data for Trends and Patterns.
- Categorize Your Target Audience Into Groups.
- Create Separate Marketing Campaigns for Each Target Group.
- Data-Driven Marketing Relies on Real-Time Insights.
Marketers are interested in three types of big data: customer, financial, and operational. Each type of data is typically obtained from different sources and stored in different locations. Customer data helps marketers understand their target audience.What are the types of data marketing? ›
Common types of Marketing data include Customer Data, Competitive Intelligence, Market Research, Commercial Transactions, Customer Feedback, Preferences and Interests, and other Marketing metrics.What are the 4 steps of data-driven decision-making? ›
- Step 1: Strategy.
- Step 2: Identify key areas.
- Step 3: Data targeting.
- Step 4: Collecting and analyzing.
- Step 5: Action Items.
The data-driven decision making approach helps organizations to formulate new products, services, workplace initiatives, and even identify trends. By investigating the historical data, companies are enabled to know what to expect in the near future and what they should change to better perform and compete.What is an example of a data-driven decision? ›
Ecommerce sites typically use data to drive profits and sales. If you've ever shopped at Amazon you have probably received a product recommendation while visiting the Amazon website or through email. This is an example of a data-driven business decision.What is driven marketing strategy? ›
A customer driven marketing strategy refers to meeting customers' needs in a more personalized way and helps businesses to optimize marketing return on investment (RoI). By identifying and targeting customers with higher lifetime value, businesses can strengthen relationships by crafting solutions to fit their wants.
What are the challenges of data driven marketing? ›
- The challenge of asking the right questions. ...
- The challenge of finding high-quality data. ...
- The challenge of breaking down silos. ...
- The challenge of normalizing your data. ...
- The challenge of interpreting your data. ...
- The challenge of working with real-time data.
It reduces the Total Cost of Ownership (TCO) by eliminating manual construction of model components by the modeler that represents products within configurator models. It reduces the time to market since the model is dynamically updated with changes in the catalog system.What are data-driven strategies? ›
When a company employs a “data-driven” approach, it means it makes strategic decisions based on data analysis and interpretation. A data-driven approach enables companies to examine and organise their data with the goal of better serving their customers and consumers.What are the 4 steps in order to designing driven marketing strategy? ›
Segmentation, targeting, differentiation, and positioning are four distinct steps that should be included in customer-driven marketing.How can I learn data-driven? ›
- Reimagine data collection.
- Ask the right questions.
- Improve your own data skills.
- Use the right tech.
- Make intelligent learning recommendations.
- Make smart internal recommendations.
- Learn from other industries.
- Use wider benchmarks.
- Predictive data analytics. Predictive analytics may be the most commonly used category of data analytics. ...
- Prescriptive data analytics. ...
- Diagnostic data analytics. ...
- Descriptive data analytics.
Most modern computer languages recognize five basic categories of data types: Integral, Floating Point, Character, Character String, and composite types, with various specific subtypes defined within each broad category.What are 10 types of data? ›
- Integer. Integer data types often represent whole numbers in programming. ...
- Character. In coding, alphabet letters denote characters. ...
- Date. This data type stores a calendar date with other programming information. ...
- Floating point (real) ...
- Long. ...
- Short. ...
- String. ...
The 7 Ps of Marketing
These seven are: product, price, promotion, place, packaging, positioning and people. As products, markets, customers and needs change rapidly, you must continually revisit these seven Ps to make sure you're on track and achieving the maximum results possible for you in today's marketplace.
- Step One: Ask The Right Questions. ...
- Step Two: Data Collection. ...
- Step Three: Data Cleaning. ...
- Step Four: Analyzing The Data. ...
- Step Five: Interpreting The Results.
Is data driven a skill? ›
Successful data scientists use data driven skills to create an emotional connection with data, empowering the business to solve the problems of customers. Data driven skills are key for Data Science experts to successfully convey facts and figures to non-technical teams.What is a data driven business model? ›
Companies with data-driven business models base their core business on data. This focus, or dependence, on data can affect all dimensions of a business model; the value proposition as well as value added or the revenue model. Value added is generated from data by making data the company's key resource.How do I create a data driven design decision? ›
- Gather enough data. ...
- Use reference points for your data. ...
- Test one variable at a time during A/B testing. ...
- Use both types of data – quantitative and qualitative, when there's a possibility. ...
- Keep in mind the context when you optimize.
As MicroStrategy reports based on results from a McKinsey Global Institute study, data-driven businesses are 20-plus times more likely to acquire new customers and six times more likely to keep them.What are the best practices for data driven decision making? ›
- Measure what matters.
- Empower your people.
- Tell stories through visualization.
- Cultivate collaboration.
- SEO. ...
- Influencer marketing. ...
- PR and affiliate marketing. ...
- Email marketing. ...
- Social media marketing.
The steps of the strategic marketing process (mission, situation analysis, marketing plan, marketing mix, and implementation and control) are different than the process for a specific marketing effort.What are the 4 common big data challenges? ›
- Storage technology to structure big data.
- Deduplication technology to get rid of extra data that is wasting space and in turn, wasting money.
- Business intelligence technology to help analyze data to discover patterns and provide insights.
- Poor Data Reliability and Scalability.
- Cost of Time and Resource.
- Blocked Projects.
- Unsupportive Service.
- Run Time Quality Issues.
Poor Quality Data
One of the big challenges of data collection is making sure that your data isn't of poor quality. Issues such as duplications, inaccuracies, lack of consistency, and a lack of completeness can all affect the quality of your data.
Which is a data driven design? ›
Data-driven design can be defined as “design that is backed by data and helps [users] understand the target audience.” It “proves that your work is on the right track… reveals the users' pain points and opportunities while unearthing new trends, and… improves your designs by adding objectivity.”What data-driven means? ›
When a company employs a “data-driven” approach, it means it makes strategic decisions based on data analysis and interpretation. A data-driven approach enables companies to examine and organise their data with the goal of better serving their customers and consumers.What are the two main types of data marketing? ›
There are two main types of data, structured and unstructured. Each contains valuable insights about your buyers. When they are combined, your marketing team can create greater context for data and expand the depth of your analysis.What is the difference between data-driven and model driven? ›
In the scientific literature, two main approaches have been proposed for control system design from data. In the “model-based” approach, a model of the system is first derived from data and then a controller is computed based on the model. In the “data-driven” approach, the controller is directly computed from data.What is driven marketing strategy? ›
A customer driven marketing strategy refers to meeting customers' needs in a more personalized way and helps businesses to optimize marketing return on investment (RoI). By identifying and targeting customers with higher lifetime value, businesses can strengthen relationships by crafting solutions to fit their wants.