Breaking Through the Complexity Barrier: One-to-One Marketing at Speed
The premise of one-to-one marketing is tantalizing: create a personalized message, value proposition or offer based on insight of customers' needs, wants or propensities. Deliver an offer just when it's needed. Say just the right words that convince customers to buy another product or accurately predict what they want when they themselves may not know. But then reality strikes. Whether trying to target hundreds or hundreds of millions of people, this theoretical level of personalization becomes virtually impossible to attain in a cost-effective, coordinated way. The culprit is a barrier created by the inherent complexity of this approach. Campaigns that begin as Formula One automobiles quickly degenerate into a mélange of bumper cars, hitting and missing their targets at random.
To address this problem, organizations turn to segmentation strategies, grouping customers so that a common value proposition may be effectively targeted and the millions of individuals reduced to a manageable number. Strategies are developed to extract the maximum value from these segments, yet the strategies typically fall short of truly realizing the potential benefit of one-to-one personalization. In an effort to surmount the complexity barrier, one-to-one marketing becomes one-to-small-group marketing. Not a bumper car, but not a Formula One racer, either. So the problem persists: how to continue to personalize at increasingly granular levels in a cost-effective and manageable way.
The answer to this quandary? Speed. Speed in data processing, in analysis, in campaign design — in any step that lies in the critical path of the creation and execution of a marketing campaign. Paradoxical though it may sound, complexity can be overcome with faster execution.
The Importance of Speed
Speed is the lever that powers the most critical component of any marketing campaign: lift. Lift is the response rate of a particular campaign. The faster the campaign cycle, the higher the lift. For large companies with tens of millions of customers, a change of one percent in an attrition-abatement campaign can deliver tens of millions of dollars to the bottom line. In an acquisition campaign, the effect can be just as large. When applied to cross-selling, increased lift will equate to higher revenues, loyalty, and profitability.
Lift, in turn, is driven by the insight derived from data. Through a variety of analytical techniques that assign propensities for specific behaviors to each customer, marketers refine their targets and seek to improve upon random chance. Lift is measured as the improvement that targeting delivers over random chance.
When Complexity Occurs
Complexity occurs when attempting to execute based on the analysis. The marketing cycle is complex. It consists of an elaborate sequence of data processing and manipulation, analysis, and the organization of scored records into campaigns that are executed across multiple channels. All this must be done before results can be analyzed and lift determined. At this point, complexity becomes overwhelming and the process becomes unmanageable as the number of campaigns increases.
Further, the length of a marketing cycle is nearly the same whether the target audience is 1,000 or 10,000,000. The only real difference lies in the counts of the data queries and extracts. The basic data manipulation and processing are the same. The analysis is usually built on a sample, so the size of the target is not really relevant and the assembly into campaigns can be complex or simple in either case.
As the groups get smaller, however, more individual campaigns must be created and executed. In other words, marketing cycles must run more frequently. For example, if a company has three segments and runs four campaigns a year to each segment, 12 campaign cycles must be executed. Assuming that each cycle takes 100 man-days to complete, you need 1,200 man-days of resources to execute. By moving only two levels deeper, you have 27 segments that you also want to run four campaigns against per year. Now the campaign cycle must be executed 108 times in the same year, with resource requirements totaling 10,800 man-days (see Figure 1).
Figure 1 — Goals and challenges of campaign cycles.
This is the crux of the complexity problem. As the number of segments grows — in typically exponential fashion — so do the resource requirements. Few economies of scale may be leveraged. In an environment where resources are usually fixed and technical and analytical resources are scarce, the complexity barrier comes up like a retaining wall during a spin-out. Management and resource constraints keep you from achieving lower levels of granularity and greater lift.
The Effect of Speed
How, then, can you reduce the resources required to execute a single campaign cycle, target deeper levels of granularity, and realize increased lift with the same fixed pool of resources? There are two ways to achieve this goal. Since you are measuring the effort in man-days, you can reduce the "man" part or the "days" part or, ideally, both. In either case, speed is the key.
If the marketing cycle is reduced to one-quarter the original duration and you hold the people constant, a single campaign goes from 100 man-days to 25 man-days. That means four times as many campaigns can be run for the same effort — four times as many segments at potentially one-quarter the size of the original. Since the segments are smaller, commonality within the segment can be leveraged more effectively to produce higher lift.
Since speed directly enables lift, speed reduces the length of the marketing cycle, and potentially reduces the number of people needed to execute a campaign. How do you achieve greater speed? By using new and emerging technologies that speed up all phases of the marketing cycle — the processing of data, the creation of models and analytical components, the assembly and optimization of campaigns, and the overall execution and management of the process.
Achieving Speed
Faster Data Processing
Speed looks different at different points in the marketing cycle. For example, the single-largest bottleneck in a marketing campaign is data processing. In no other area does speed have as much of an effect. In an analytical effort, 80 percent of the time is typically spent on data cleansing, enhancing, and transformation. In addition, data freshness can be impaired if refresh cycles are limited by data loading times. Even if no other improvements are made, reducing the duration of this step can significantly reduce marketing cycle time.
To achieve this improvement, new technologies can be leveraged to cleanse, house, de-dupe, and match data. These new technologies can increase the data refinement and loading rate from 100 gigabytes per hour to one terabyte per hour. That means data loading times can be reduced from days to hours. And the time it takes to create analytical files can be reduced from days to minutes. Business analysis can be done in real time, not in daily cycles. List generation can require only the time it takes to write to tape. These technologies exist today and can be the key enablers to reduce data processing time.
Faster Modeling and Analytics
Emerging technologies also can be used to shorten the analytical process. New analytical and modeling tools can reduce the creation time for an analytical model from four weeks to hours or even minutes. In addition, many of these tools are automated, which reduces the "days" part of the metric as well as the "man" part. Are the models as effective as those that take four weeks to build? As with all work of this kind, it depends on the user. All things being equal, however, these fast, automated tools have produced models that are at least as powerful as the ones built manually, if not more so.
Faster Campaign Configuration
The next item on the critical path is the configuration and optimization of campaigns. This is a particularly difficult step. Not only does it require a great deal of planning and organization, it also requires the marketer — traditionally, a non-quantitative person — to deftly juggle priorities, dollars, and offers to achieve the best possible ROI.
Typically, this step is performed manually and motivates the initial segmentation. By creating groupings of customers that represent a common level of value or set of interests, a marketer can take a common approach to the group and simplify complexity. While this approach is better than sending everyone the same offer, it still fails to maximize ROI.
True optimization requires that the cost and profitability of offers and customers be analyzed in the context of each individual's propensity to buy. Then the optimal match of customers to offers must be calculated. Naturally, this approach can lead to tens of millions of possible combinations and is virtually impossible to execute manually. However, new technologies can automate this entire step and optimize ROI, response rate or number of acquisitions. Importantly, all of this can be done in a matter of hours rather than the days or weeks that is typical now.
Faster Execution
All these new technologies help shorten the individual steps needed to create and execute a campaign. Collectively, they can shorten cycle times significantly and deliver the associated benefits of fewer resource requirements, more time for strategic thinking and increased lift. However, these individual advances still leave a very complex process to orchestrate across disparate and sometimes antagonistic departments.
That's where workflow technology comes in. Recent advances make process facilitation and automation viable in marketing departments. New technologies allow managers to define the steps their departments will take to execute a marketing cycle, as well as define the timing, the resources responsible, and the contingencies. Managers also can use the same tools to monitor the execution of the process, which allow them to see bottlenecks, resource constraints, and potential improvements that can be incorporated into the next cycle.
Workflow technologies may seem antithetical to marketing departments, because marketing has traditionally been the "creative," less structured side of an organization. However, marketing has evolved into a discipline that requires science and rigor as well. Given that marketing is now responsible for managing what may be considered an organization's greatest asset — the customer relationship — it should be no surprise that companies are looking to rigorous integration and management tools to more effectively optimize customer interactions.
Process facilitation workflow accomplishes this goal. As the processes are to be executed, the resources responsible for a particular process step will be alerted to the "job" in their in-box. This job will contain all of the relevant information, such as a description and due date, as well as notices of where to send results, what to do in case of problems, required levels of quality, and exit criteria.
This level of integration and communication automates the hand-offs that are the traditional log jam in business processes. Workflow technologies liberate the team to focus more on specific tasks and less on the coordination of tasks between people or departments. In addition, simple or repeated steps can be programmed and automated to further reduce the time and effort necessary to execute the process. Figure 2 illustrates process facilitation workflow in action.
Figure 2 — Process facilitation workflow
Workflow management takes on even greater significance today given the virtual nature of teams. While a process may be clearly defined, different parts may be executed across the country or even around the world. Add to this the probability that part of the team may be a subcontractor or consultant and marketing complexity becomes even greater. Process facilitation workflow enables marketing managers to define and manage a process across the virtual team just as easily as if the team were co-located. The technology works just as well over distributed networks as it does within a single building.
Making It Work
Even sophisticated marketing departments may look at this list of improvements and find it overwhelming. Some clients have even claimed that such a level of improvement would create chaos in their organizations and is too risky. It is true that significant transformation is both risky and challenging. However, when just one percent of lift can deliver tens of millions to the bottom line, the business case is compelling.
Given this value proposition, the challenge becomes how to find the lowest-risk way to implement the change. The first step is to derive and define requirements directly from a well-articulated customer strategy. Once articulated, these requirements can be satisfied in a number of modes. An in-house solution allows a company to exercise complete control over the effort and reduce risk, but it may disrupt current operations. If the impact on current operations must be minimized, an outsourced model that relies on well-defined service level agreements can be the appropriate alternative. Typically, however, a combination of these two modes is the best solution to reduce risk and maximize benefit.
In and of itself, speed does not need to create chaos if the right tools, systems and capabilities are in place. Formula One automobiles use the best electronics, the best fuel and the best drivers, and they run the fastest races in the world.
At a time when marketing to tighter and tighter target markets is a complex task of gargantuan proportions and compelling necessity speed is the solution, not the problem. And now, thanks to new and emerging technologies, it is possible to man-age large data in real time, automate and enhance the modeling process, optimize and configure campaigns in minutes, and facilitate the definition, execution, and manage-ment of the overall marketing process. While the implied change can be difficult to embrace, the known benefits are too compelling to ignore.
