Building Absorptive Capacity

On the one hand, implementing an innovation is a project. You know when the preparatory work is to be done, when the cutover will be made, and when the system will be humming along. The change needed to be able to absorb multiple innovations, on the other hand, is much more than a project. The only problem is the executive who sees a Big Data investment as a project. We know that executive sponsorship is important, but frankly, executive sponsorship can't be limited to an exec who just says, “Do it.” I've seen million-dollar systems (at least in terms of total costs) fail completely because the exec said to do it but then handed everything off. If the exec won't use it, then no one will. That's taking a project management–only approach.

Further, with Big Data and DCS, we're really asking for a lot of things to change at the same time. So we just implement standard change management, right? So, at least from the change management side, you have to get buy-in from the rank and file. When you're working with merchandisers or salespeople, they're the ones you have to convince. But it's often the case that their benefits from whatever you're doing are either far off in the future or just simply marginal to some of the users, and those are users who are critically important to the overall success of the full system.

It's no wonder, then, that we heard a decade ago that up to 70 percent of CRM installations were failures.⁵ At least one reason was that it was just too hard to separate change from the project. And while we're not hearing that kind of failure rate number today, I think we're only just beginning to understand what happens at the intersection of change and project management. Change isn't a project. Of that, I'm sure.

The answer lies in building absorptive capacity. Absorptive capacity, that ability to learn and master innovation, technology, and change, is a competency, something that some organizations do better than others. The good news is that you can build this competency.

The rest of this chapter is devoted to the adoption of Big Data technology and building absorptive capacity. The next chapter then covers the cultural shift necessary to fully embrace Dynamic Customer Strategy in order to get the most out of the Big Data investment.

People, Process, and Tools

Building absorptive capacity requires making the right changes in people, process, and tools. There are many who talk about people, process, and technology (or tools) around other aspects of business, and absorptive capacity is no different. There is an intentional element to absorptive capacity; it doesn't have to be left to chance. These intentional strategies designed to increase absorptive capacity have to address people, process, and tools.

People

McKinsey Global Institute reported in 2011 that another 150 million data-capable businesspeople are needed if businesses are to capitalize on Big Data.⁶ Further, many organizations say they currently lack the necessary analytical skills, such as predictive model development, data (or text) mining, and overall statistical analysis skills, for successful data discovery.⁷ If these studies accurately reflect the state of today's Big Data world, then finding and successfully recruiting people who can contribute to a customer knowledge competence (CKC) is not going to be easy. They are in high demand and scarce now—and the situation will only get worse.

The process of recruiting, selecting, and developing competent people is, in the general sense, well documented. But we need 150 million and the skills and knowledge that they have to have aren't universal. Some will need to be statisticians; others must be marketers; still others merchants and salespeople. All must have the right set of Big Data and DCS skills for their position. The challenge associated with the people dimension is that Big Data and CKC are not just based on one technology. Many critics and experts want to view Big Data as yet another innovation that will go through the standard adoption curve. In truth, the effective use of Big Data requires more than skilled data scientists; it also requires strategists who understand how best to deploy the tool and tacticians who organize and manipulate data into operational models.

Further, just having statisticians isn't sufficient. There are many examples of organizations that are great at using Big Data for various purposes that never enter the realm of customer acquisition → retention → growth. Volvo, for example, does a great job of using data to build safer cars, but until they can also use that data to build exciting cars, they're bound to be a minor player in the market.

Figure 10.3 illustrates that customer knowledge competency is based on a blend of technology, statistics, and marketing. That means you have to seek out people with at least two of the skillsets. Or you have to think about how you'll develop them from one to two, then three. Given that most data scientists today fell into their competency area (i.e., marketing, finance, security, or whatever), I suspect that it is just as easy to make one as it is to find one ready to go. And it is easier to add marketing to a data skillset than the other way around, though not impossible.

Note that all must understand marketing. Maybe I'm overemphasizing the point, but a Dynamic Customer Strategy requires that everyone understand the consumer, which means they have to also be data savvy.

Just as important, the knowledge has to be as vertical in the organization as it is horizontal. In other words, management has to interact with data scientists.

Starbucks is a great brand. But their CEO doesn't understand data and doesn't interact with data scientists.⁸ Neither does the rest of the executive team. That's why they made those horrible blunders: overexpanded sites, overexpanded product categories, and other poor decisions. And while the current CEO may be effective at shrinking the company back to profitability, they aren't leveraging their customer data effectively because management can't be bothered to develop a comfort with customer data. They're competent, just not customer knowledge competent, and that will limit their ability to grow profitably.

Contrast that with Cabela's. Cabela's is a publicly traded company, just as Starbucks is, but its board of directors regularly receives presentations from data scientists. Cabela's board is as familiar with the company's customer personas as are the executives, merchants, and marketing staff. As a result, there is a culture that epitomizes the principles of Dynamic Customer Strategy based on Big (Customer) Data.

While strategy and tactics are necessary components of an effective CKC, there are other people needed for other tasks. Yet CKC involves many tools: statistics (of which predictive analytics is but one form), information technology (Web, mobile, etc.), enterprise data warehousing, and marketing research (both qualitative and quantitative, which require different skillsets). Further, since CKC is neither just the acquisition of customer data nor the interpretation of that data but the integration of customer insight into decision making, process engineering is a necessary technology for effective CKC.

Managing the Change

One important factor in getting a Big Data proposal accepted and implemented is executive sponsorship. Someone at the top has to lead the charge, because adoption of Big Data technology is more than just a simple purchase of software. Executive sponsorship is necessary, not just for getting the proposal accepted but also for getting it fully implemented, but executive sponsorship during purchase is, by itself, insufficient, as you can see in Figure 10.4.

Currently, we know of one CRM manager in a B2B division of a major company who is afraid of losing her job because she can't perform (that's why we can't use her name or the company name). But the reason she can't perform isn't that she's incompetent. Rather, the problem is that the division CEO mandated the purchase of a CRM software system and hired her to run it, but that was it. He took no further action to support what was, at least to him, a project to get done that quarter. Then he left and the replacement didn't care. Salespeople who had to enter the necessary customer data to really power her solutions didn't cooperate. They couldn't see any personal value so they didn't enter the data; as a result, the lead generation solutions she came up with were less effective—so ineffective, in fact, that the salespeople wouldn't use the leads. If they don't call on the leads, the leads don't become customers and the whole thing spirals into a black hole. Since she's afraid of the negative publicity, we can't share the name of the company, but I'm certain you have some of their consumer products in your home right now.^*

When faced with adopting a system that will be used by a wide array of users, several important lessons can be learned from this example. The first is that the executive sponsorship has to be about the change, not the project. When you adopt Big Data DCS, you're not buying a product—you're changing the way people think about strategy and about data. Big Data DCS can't be the focus of a single quarter. I hear a lot of leaders in the more data-mature organizations talk about their Big Data “journey,” because it really is a journey. What you adopt today is but a single step along that way.

Another lesson from that example is that one important set of users, salespeople, didn't see the personal value so they failed to comply. One reason they couldn't see the value was that they didn't participate in the selection process, the design process, or the implementation process. As a result, a system was bought, built, and implemented that didn't deliver much value to them. Yes, they'd get better-quality leads, but that's only a small part of the potential value, and the return on leads is historically so dismal that a new system isn't going to sell itself. Involve users in the selection, design, and implementation (as illustrated in Figure 10.5). You don't have to involve all of them, but do have representatives that the rest of the users know and trust.

The third lesson here is that they failed to comply because they were allowed to not comply. They didn't have to so they didn't. If you need the cooperation and compliance of users for the system to work, the system has to be necessary for them to do their work.

Did you see the two things going on here? First, if a system is going to be necessary to do the job, it has to fit the job that needs to be done. That means that you select technology that works with your work processes. If you have a sales process, the technology should support it—as Ram Ramamurthy, one of my colleagues, says, don't cut off your foot to make the shoe fit.

The second thing is that if you don't have a sales process, you better figure that out before you buy the technology or you'll force your salespeople into a process based on the technology, and that is always a recipe for disaster. You have to have processes first before you can automate them.

Another possibility is that you have a sales process but the system isn't necessary for the process. If the data is to support marketing but salespeople get along quite successfully without using the system before the sale, then marketing has to get the data elsewhere. Such is the case at AT&T. Their salespeople don't use the system except to input orders, so it is effective at managing customers, but not prospects and other noncustomers.

Further, you must eliminate workarounds. For example, salespeople have to complete and turn in reports regarding who they called on and what was the result, sales forecasts, and the like. A good CRM system eliminates those reports and allows the manager to pull the data directly from the system. When Konica-Minolta Business Solutions (KMBS) adopted CRM technology, salespeople were allowed to make changes to the data that essentially allowed them to configure their own territory. That's not healthy. Further, the system rolled up only 90-day forecasts in an environment that had an average sales cycle of two weeks. As a result, the forecast was useless and sales managers and salespeople had to work around it. The workaround was to manually compile forecasts on a weekly basis in keeping with the short sales cycle, and many salespeople either didn't turn one in or turned it in late, causing problems in inventory management. Now the company is in the process of changing vendors so that they can get a system that fits their work processes more completely and so they can eliminate workarounds.

Similarly, make sure that users' metrics and evaluation/reward structures align with successful use of the system. If people are rewarded to do something different, that's what they'll do. If your evaluation and reward structure doesn't align with the system's intended use, why are we doing this again?

Pilot the system. First, a good pilot eliminates problems before it reaches the field. If a lot of users are going to be on the system, test and make sure all of the bugs are out before the system goes live for everyone. Most vendors will allow you to pilot the system anyway before you buy it, and the pilot will enable you to validate that you'll generate the necessary ROI to make the purchase worthwhile.

Generate quick successes. During your design phase, you've identified potential power users, those merchants or salespeople or marketing people who are likely to be early adopters or innovators, and you invited them to help design the system. Working with them, ensure that some quick successes happen in order to generate positive word of mouth among the users. Gallery Furniture's adoption of a CRM system for its sales staff led to two salespeople doubling their sales in only 30 days. You can try for that kind of overall early success or use the big-win strategy (“Joe here won the big deal at Mega-Co thanks to his Twitter strategy!”). Do both if you can—people respond to stories, but you have to have the stories first.

Finally, when rolling out the full system to the field, carefully plan your launch strategy. Your launch strategy should be comprised of three components: communication, training, and reward/punishment. The communication component should include messaging around why the system is important for the company, how it will benefit the user, and information on training and support. If you want buy-in, you need a communication strategy that sells.

You also need training and support commensurate with the complexity of the system. Simple systems require less training and support; complex systems need more. Obvious, I know, but remember, users decide what is simple and what is complex, not your IT department.

As part of your launch strategy, you should also consider a series of early rewards that encourage trial. For example, you could have a contest with a small prize available to every user who completes five sessions in the software. Whatever it is, use rewards to encourage trial, not success. Success should already be rewarded in the normal evaluation and reward system—for example, salespeople get paid to sell. If the system helps them sell, they've already gotten that reward. Focus instead on the desired activity, such as entering data. Similarly, and tied to eliminating workarounds, devise appropriate punishments if they fail to comply. If someone hasn't logged into the system in the first three days, for example, perhaps a call from a manager to explain why may be enough.

Table 10.1 summarizes all of the elements you should apply to secure a successful Big Data implementation. What do you do if you don't have control over all of these elements? Most of the time, a Big Data play involves multiple functional areas, areas that you probably don't have control over. You may not be able to manipulate a compensation or evaluation system, for example, but have to work within one that is managed by someone else. If that's the case—if there are key components that are going to be managed by leaders in other areas—you've got to have them on board from the beginning. If they refuse to change the compensation structure, for example, your challenge lies not in making a Big Data play but in understanding what is important and why the company is paying for that activity and not the type of activity you're trying to accomplish. Ask questions such as, “Why are you willing to budget x dollars for this Big Data application when it won't drive the performance you measure and reward?” If strategic objectives don't align, you're fighting an uphill battle.

Empowering Your Entrepreneurs

Stop for a moment and put the book down. I know that's hard to do, but try. Think back to how your company reached out to customers five years ago. What was the customer experience like? What were the strategic points of emphasis for your company?

Now think about what your organization might look like if you could rewrite the history of your company for the past five years if the ideas and principles of everything you've read so far had been adopted back then.

Would it look different?

Of course, and probably much different.

For one thing, you can look back at missed opportunities and see how those might have been captured through application of DCS. You may also realize how your company could have failed quicker because you would have had greater control in experiments to identify causality, and moved on to better opportunities rather than agonizing over the lingering death of a doomed program.

Together, these actions of identifying opportunities and responding to them are the basic activities of the entrepreneur. Yes, there are others, like assembling capital, but at least according to the entrepreneurship scholars, these are the two fundamental activities of entrepreneurs.

Big Data and DCS empower the entrepreneurs in your organization to sense opportunities—or to control them. I especially see this tension in the field: Systems get put into place to automate reporting up rather than offering insight down.

In our study of sales execs, this dichotomy was obvious.¹³ Few, very few execs were talking about how technology could be used to drive information to the sales reps and empower them to make better decisions about their customers, their territory, and their sales strategies. Most talked about the value of the system for reporting activity and providing greater transparency of salespeople's activities, giving managers an opportunity to manage more closely.

Similarly, we observed the same dichotomy in our study of retailing data maturity. The more mature organizations are thinking of how to make technology empowering rather than controlling. These leaders were trying self-serve analytics so that merchants and store managers could make their own decisions. One best practice we've seen is to create a reporting exchange—a place where users can showcase their work and other users can adopt or tweak it. If you have self-serve analytics, then users will want to use those tools. Speed up the innovation process by giving them a place to share what they've learned about how to use the tools.

I've had several retired JC Penney executives or store managers tell me that the end for JC Penney began when the company took inventory management away from the store manager. Local product selection was a keystone of Penney's, but in a move to become more efficient, product mix decisions were centralized. The result was less attractive assortments that led to the need for more regular discounting, which then trained buyers to expect everything to be heavily discounted. That was the problem that the company was trying to fix when it began that disastrous experiment with fixed pricing.

The counterargument, of course, is that this decision was pre–Big Data and that with today's data we should be able to develop better predictive models that should enable better centralized decision making. That argument is correct. We should—up to a point. But at some point, you really have to question whether you need a store manager or a salesperson in the field. If this person isn't able to add value, then why is she there? Predictive models, no matter how good, need some human insight, and that is best done by the person closest to the customer. Use data to empower, not only to control, so that your people can use their entrepreneurial talents.

And yes, not all are entrepreneurs. Some will use the information only to self-manage, and that's fine. If ever the Pareto principle (80/20 rule) was true, it is true with entrepreneurs. You only need a few to experience awesome results.

Konica-Minolta's Awesome Results

Konica-Minolta Business Solutions (KMBS) used to be a copier company, but that industry is really all but gone. Only a few years ago, there were 20-plus manufacturers, now there are four or five, and that number may be smaller by the time you read this.

What was certainly a crisis to many former competitors was simply an exercise in pivoting—in recognizing the opportunity and pivoting to meet it. A critical element in that pivot—from hardware-pushing copier company to solution-focused information technology provider—was the approach KMBS took to empowering their salespeople with data.

“Our first CRM solution was really more about controlling salespeople rather than empowering them,” according to Velinda Cox, KMBS vice president. “In fact, it was almost punitive because it focused them on inputting activity rather than the richer intellectual property you really need to manage quality relationships.” Now the company is making the transition to its third-generation CRM solution.

“Our system integrates multiple data sources, including servicing, billing, and logistics.” KMBS also sources third-party data so account execs can build a deeper understanding of their accounts.

Now, instead of focusing on closing a deal, the sales exec works to build a plan with the account that seeks to accomplish the account's strategic objectives. Surprisingly, sales cycles have actually gotten shorter, with an average sales size that is three to four times that before the change.

“Before they (salespeople) were given a phone book to find customers, today we identify potential accounts based on “propensity-to-win” models and a propensity of need—using Dunn & Bradstreet for firmagraphics, what have they already bought, our transaction data, and our survey data (customer voice) that helps us identify what we do well and what we don't do well and use that to get better.”

What she describes is empowerment. And in case you missed it, shorter sales cycles and three to four times larger average sales are pretty awesome outcomes.

One Result: Customer Knowledge Competence

The KMBS story illustrates how real-time insight from real-time data—using the velocity and variety of data—can be one benefit of Big Data. Without systems that bring us data as it happens from multiple sources, we can only describe history, and frankly, the better we are at creating systems that describe history, the worse we become at anticipating the future.

It's worth repeating the quote from Jack Welch: “The only sustainable competitive advantage is to be able to learn faster than your competition, and to be able to act on that learning.” Customer knowledge competence (CKC) is exactly that ability—the ability of the organization to acquire, analyze, disseminate, and act on customer information.¹⁴

Yes, the book up to this point has clearly shown you how to build a customer knowledge competence. So what else could possibly be needed?

CKC is like any form of organizational competence. For the competence to develop at the organizational level, three things are needed: people with the right skills and attitudes, processes that support the development and execution of the competence, and tools for use in the execution of the competence (see Table 10.2).

Source: John F. Tanner Jr., Dynamic Customer Strategy: Today's CRM (New York: Business Expert Press, 2013), used with permission.

Recall that we said in the first chapter that one barrier to Big Data and DCS is a fear of math. Further, we also noted that there's a huge shortage in data-capable managers. Whether you plan to develop human resources internally or find them in the market, we have to recognize that right now, we're limited in what we can do in the people area. In fact, we found in our data maturity study that a key activity of the customer insights group was simply teaching merchants and sales managers how to use data. (We're trying to do our part in colleges and universities, but that takes time.)

Tied into the development of people is the deployment of processes and policies that support a CKC. Note the subtle difference. I'm not talking about the processes that flow from a CKC, or those processes that were the subject of the book so far. Rather, I'm talking about processes like reward systems that make this competence desirable and policies that make professional development into DCS and Big Data possible. In one organization, our research team observed two prominent displays in a call center: a sign saying “Customer Satisfaction is Job 1” and a “clock” displaying the average call time. Management may have wanted customer satisfaction, but they were rewarding shorter calls, so shorter calls were what they got. If you want CKC, you have to reward it, and if you want professional development so that your people can use data more efficiently, you have to make it possible.

Finally, you have to equip your people with the right tools so that information can be converted into knowledge. I use the word “tools” and not “technology” for a reason. Technology may be an important component, such as data warehousing technology, marketing automation software, and the like, but you also need include data acquisition tools such as surveys and focus groups and statistical tools such as hierarchical linear regression or cluster analysis. Further, you need reporting tools. But even your concept map is a tool if you use it that way. Your map becomes a teaching tool for explaining strategy, a planning tool for guiding operational choices, and a measurement tool so you can see how you've progressed.

In other words, I expect you to use your map as more than a residence hall for dust bunnies.

Global Implementation

What happens when these systems must be implemented globally? Obvious challenges, such as language, can impact the success of a global implementation. You may do business internally in English, but local laws may require that the software be translated into the local language. And language is but one challenge.

You may have a company with a single corporate language, such as English. That doesn't mean that users in various countries understand English well enough to use only English training materials. Consider the language needs of the users when developing training materials and training programs (such as online tutorials). Even if the law doesn't require translating that software into the local language, don't overlook the importance of local-language training materials and programs.

What happens to service? Will it be available 24/7 and will it be available in the local language? Scandinavians work in the wee hours of the morning during fall and spring in order to make the most of what limited daylight there is in the afternoon. Latin Americans, however, tend to start late and stay late.

And don't forget the data. In Europe, laws are much stricter on what is permissible use of data, even machine data. Germany, for example, requires a double opt-in process for any personal marketing, and many European countries do not allow for transfer of data out of the country. That can make global account planning very difficult.

Those are the obvious questions: language, service, and data and privacy laws. But there are other considerations as well, depending on the nature of the tool.¹⁵ For example, if you are implementing a new business intelligence tool, does it have to manage currency differences? Local users may need reports in local currency, while corporate users need a single currency for roll-up.

Global implementation is often a top-down process. If users are involved, companies tend to only involve those users within or close to corporate headquarters. Compliance can vary tremendously as a consequence, and with possibly huge implications. If your returns are predicated on global usage, you may need to involve global users in the design process. I suspect, too, that you'll learn a lot about how things can be done because you'll observe greater variance in how work is successfully accomplished by involving them in this process.

Whether you did or didn't go local in the design process, you'll have communication challenges when you launch. Clear communication in local languages of the reasons for the new tool, benefits to users, and so forth is just as needed as it is when you are addressing only a domestic launch. Nikolaus Kimla, CEO and cofounder of Pipeliner Sales, says this is one place where your vendor should help. If your vendor has local case studies or reference clients, use these to communicate value in each country. (If your vendor doesn't, do you have the right vendor?)

Decentralization of leadership can also create challenges. We mentioned KMBS earlier—their CEO is in Japan but the company's largest sales organization is in the United States. Just getting virtual meetings coordinated to make sure the launch will go smoothly is going to be a challenge and will slow down everything.

Summary

Big Data requires the adoption of many innovations. The technology needed to capture Web browsing data on your website, mobile technology for data capture and push messaging, automated Twitter drip campaigns, and lots, lots more are all part of the Big Data revolution. Building absorptive capacity is needed to accelerate the adoption of these innovations nearly simultaneously.

The objective is to create competency in customer knowledge—to be competent in applying Big Data to create streaming customer insight. To be competent requires people who can use tools effectively and apply processes to continuously get better. In other words, building absorptive capacity and customer knowledge competence can be accomplished through intentional actions.