The Better Half: The Artful Science of ROI Marketing
Advances in analytics, systems, processes, and organization design can converge to eliminate the waste from marketing.
(originally published by Booz & Company)
That uneasy sense of mystery has undergirded the growth of modern marketing communications. What “works” best — sustained brand advertising, or targeted retail promotions? Is “likeable” media advertising a more effective vehicle than hard-selling direct marketing? Can trade promotions help boost share permanently — or do they destroy brand value? Because companies believe they cannot tell which portion of their marketing spend leads to steady increases in sales, share, and profitability, they rely on imperfect metrics and anecdotes to guide their marketing programs, which today account for hundreds of billions of dollars in worldwide spending annually.
The public has certainly benefited from businesses’ confusion about marketing effectiveness. Each year, marketers’ “generosity” creates thousands of new products; funds hundreds of new advertising-supported “free” media, including magazines, broadcast networks, cable channels, and Web sites; and supports millions of jobs in advertising, marketing, and sales.
Unfortunately, marketers themselves are not gaining. Marketing activities constitute a rapidly growing portion of companies’ cost structure. Meanwhile, the returns are growing even more doubtful:
- Advertising and media, trade promotion, and consumer promotion spending now account for as much as 20 to 25 percent of sales among consumer packaged-goods companies in the United States, up from 15 percent in 1978. Trade spending today is the second-largest item on the profit and loss statement for most of these companies (following cost of goods sold).
- Advertising spending continues to balloon, despite audience fragmentation that is rendering it less and less efficient. Last June, the major U.S. broadcast television networks took in a record $8 billion in upfront advertising revenue, a $1 billion increase over the previous year, even though their audience share against cable continues to decline, to 37 percent in the summer of 2002 from 41 percent in 2001.
- The return on marketing spending, historically quite low, is going lower. At the Big Three automotive manufacturers, incentives have more than tripled since 1990, reaching nearly $3,800 per vehicle, or 14 percent of the average sales price, according to CNW Marketing Research. Yet Detroit has continued to lose share in the U.S. (by 1.6 percentage points in 2002 alone) to imports whose incentives are half as high.
Marketers typically rationalize this lack of productivity with the modern-day equivalents of the Wanamaker wisecrack. “Well, at least we’re building awareness,” they say — rarely stopping to explain how, if at all, attitude changes translate into actual, lasting consumer buying behavior. “We must support our brands” — not knowing whether that support does anything other than make the brand famous. Pressed, they will concede the gap in their approach to marketing. “Are we underspending and leaving money on the table, or should we spend more, and if so, on what?” they will ask. “No one really knows.”
Wrong. Companies can know where and how to apply marketing expenditures to achieve significant, lasting lifts in a product’s or service’s profitability. You might call our approach “Wanamaker’s Revenge.” We call it ROI marketing.
ROI marketing is the application of modern measurement technologies and contemporary organizational design to understand, quantify, and optimize marketing spending. The result: improved return on marketing investment, achieved through analytics-based decision making that directs the funds toward the executions, pricing, product adaptations, vehicles, and/or territories where they will generate sales more profitably.
All marketing activities are amenable to ROI measurement and optimization: trade and consumer promotions, pricing, media advertising, product placements, even product assortment and the content of sales calls. Brand marketing campaigns can be assessed on the basis of returns to the bottom line. Since we began developing and fine-tuning our frameworks in the late 1980s, we have used ROI marketing analytics and programs to help a leading frozen-food manufacturer realize 10 percent sales growth on an increase of less than 2 percent in trade spending. A pharmaceutical client saw a nearly 5 percent increase in operating profitability over a one-year period after it reallocated resources for medical education, detailing, and patient support services across physician and patient segments. And by determining the true value of individual components in a media distributor’s merchandise assortment on the basis of an analysis of expected sales, gross unit margins, and return costs, we enabled the company to modify its product lineup and quadruple sales of selected products over a short period of time.
ROI marketing is not a fix, but a philosophy. Typically, companies approach the marketing predicament from an isolated vantage point: It’s a problem of analytics (hence the quest for ever-better modeling tools) or systems (thus the growth of the CRM industry); or it can be solved by better processes (we can fix it at an offsite meeting!) or organizational changes (such as getting sales and marketing to cooperate once and for all). ROI marketing acknowledges that improvements in the effectiveness of marketing must be all-encompassing.
Such comprehensiveness, of course, requires a fundamental transformation of the organization, led by the CEO and his or her executive team. Transforming a multidivisional organization in order to optimize marketing spend can be enormously difficult. Decision rights between sales and marketing must be reframed, with less micromanagement by marketing and more authority and accountability flowing to sales. Information technology systems frequently must be rebuilt; often the supply chain has to be reconfigured to harmonize product development, distribution, and marketing communications.
But the hardest part of institutionalizing ROI marketing is changing the century-old mind-set that posits, often explicitly, that marketing is solely an art, measurable only on the basis of customer attitudes, however tenuously they link to consumers’ activities. Installing a foundation of fact inside the marketing organization, then rebuilding behaviors and relations atop it, all toward the goal of creating a profitable, durable company is the greatest challenge senior executives will face.
We are not claiming that marketing can or should become the stepchild of analysis. ROI marketing cannot substitute for the creative spark that ignites the best new product ideas and ad campaigns. But an ROI marketing capability can accomplish something companies have sought vainly since the dawn of modern marketing communications: accountability attached to marketing creativity, which will allow executives to find the half of their marketing spend that really works.
The Case for Change
“American advertisers rely on ‘essentially illogical’ approaches to determine their advertising budgets,” wrote University of California sociologist Michael Schudson in his 1984 book, Advertising, The Uneasy Persuasion: Its Dubious Impact on American Society (Basic Books).
Illogical, but not irrational. Marketers’ increased reliance on promotions, pricing, and big blasts of media advertising to boost volume is a lucid response to the pressures they face in mature markets. Wall Street typically uses volume growth as a proxy for the underlying health of a business, especially in consumer products and other mature industries. The Street reasons that sales increases indicate a reliable, sustainable, and growing cash flow. Hence, companies that meet volume targets frequently are rewarded with rising share prices and healthy multiples.
So important is volume that companies frequently will depress profitability to sustain volume. Indeed, battling for volume using pricing, promotions, and other primitive instruments is like the Cold War: If the volume growth isn’t profitable, the increased “arms spending” can damage, even kill, the enterprise. When Philip Morris Companies Inc., concerned about Marlboro’s eroding market share, slashed cigarette prices in April 1993, it added five share points within eight months. But the profitability of its cigarette business took more than five years to recover.
Senior executives, by and large, are aware of the dilemma but have felt powerless in the face of growing retailer dominance. According to both our experience and analyses done by the companies themselves, trade spending by major consumer packaged-goods manufacturers has an average ROI of negative 20 percent — that is, for every dollar spent to generate incremental volume, marketers gain a short-term return of 80 cents. Retailers, by contrast, see profits from almost three-quarters of all promotional events.
Worse yet, this Cold War is being fought inside companies themselves. Firms are typically driven by the competing objectives of the marketing and sales departments. Sales personnel are rewarded for volume increases, yet marketing frequently owns the P&L. At many companies, an “iron curtain” goes up between these units, with marketing executives centralizing authority in a vain attempt to control and improve sales promotion and local marketing; meanwhile, out in the “real world,” sales pushes volume at any cost. Linkages between brand strategy and sales promotion plans are not fully considered, and account-specific profit objectives are deemed a pipe dream, too difficult to set and track. The constant bickering leads to a cultural stalemate: Marketing executives are considered the brains, while sales executives are considered the brawn (or salespeople are thought of as the pragmatists, and marketers the pointy-headed intellectuals). Wherever the truth lies, the concept of working together toward a common goal is lost among the battles for budget and primacy.
A case for change, then, must be built both within the marketing enterprise and among its significant constituencies, particularly retailers and investors, to gradually wean them from their dependence on volume and redirect them toward profitable volume. Probably the best way to catalyze such a case is to start small, by picking specific targets, identifying events that can be rendered more profitable, and then showing the profitability improvement.
An event — the “atom” of ROI marketing — is a communication through a vehicle for a period of time. Just as the visible universe consists of countless combinations of the 112 identified atoms, the modern consumer economy might well be viewed as billions of events and their outcomes: a coupon in the local newspaper offering 15 cents off a well-known toothpaste this week only … a sales rep’s call on a doctor to leave behind a sample of a new pharmaceutical product … a one-month network advertising flight supporting the introduction of a new automotive marque. At the end of the event period, the marketer should know whether it achieved its goal.
Consider a typical sales promotion scenario: a special display in 15 percent of a grocer’s stores in a given metropolitan statistical area (MSA) promoting Kleenest Detergent in the Power-Wash Box at a price of two for $4 for two weeks, for which the manufacturer provided a $5,000 lump sum payment to the retailer. At the end of that period, both retailer and manufacturer will understand whether the event was profitable for them, which will provide a basis for future negotiations and decisions.
Well, that’s the theory. In fact, event decisions like that, determined territory by territory, product group by product group, week in and week out, by busy sales reps and harried retailers, are almost always imperfectly grounded and subjective. If Kleenest Detergent gets a great lift, is that lift owed to the price, the display, or the fact that the event coincided with the July 4th holiday? That brand advertising flight: Did it drive our volume uptick — or did the competitor’s price increase? Can we make a causal linkage between the advertising and sales? What made the difference — increased media weight, or an individual commercial?
Hard to tell. The tools used by sales reps, store buyers, media planners, and marketing executives are simply not robust enough to support the on-the-fly decisions they face day after day.
Hence the need to create a capability inside the company. We define capability as the cross-functional, intra-unit capacity to do something better than before, and to do it consistently, inside the organization. An ROI marketing capability comprises gathering and evaluating data, making the right decisions, implementing them, measuring their impact, and finally adapting the organization based on the outcome, all toward the goal of achieving more profitable sales.
To make this more concrete, we break down the ROI marketing capability framework into four components. These are:
- An Analytical Engine. An intricate statistical calculation model, the analytical engine enables a better understanding of the incremental volume different events generate, clearly identifying profitable events and helping companies avoid unwittingly planning events that produce negative returns.
- Decision-Support Systems. Complex expertise built into user-friendly (albeit sophisticated) tools allows personnel to collect, integrate, and apply data from the analytical engine and the field to support actual business activities, such as planning advertising or promotion events, measuring results against plans, etc.
- Redesigned Business Processes. Business strategy and budgeting, target setting, tactical planning, and post-event analysis must be coordinated and embedded in the organization to achieve the right profit, volume, and spending targets event by event, account by account, geography by geography.
- Organizational Alignment. Establishing clear decision rights, training and empowering executives and staff, and developing appropriate incentives all help decentralize tactical, knowledge-based decision making within the context of a corporate strategy newly balanced between volume and profit objectives.
We stress the importance of capability building and organizational transformation in part to distinguish ROI marketing from the rhetoric of CRM systems enthusiasts, who tend to equate technology with solutions. Whether the customer is the consumer or a business-to-business buyer, customer relationship management is fundamentally a human activity; technology can aid it, but it cannot substitute for it.
Exploring this capability framework in more detail can help reveal the extent of the transformation, and the returns that accrue from it.
Analytical Engine
“What’s the difference between Giant and Jumbo? Quart and full quart? Two-ounce and big two-ounce? What does Extra Long mean? What’s a tall 24-inches? And what busy shopper can tell?” the writer Marya Mannes once asked. Little did she understand: They are different — and are bought in different ways by different consumers in different places at different times for different reasons. But distinguishing among the array of reasons and building concrete marketing plans based on that knowledge has been well nigh impossible.
A company could draw a statistical model showing how sales of one stock-keeping unit (SKU) — that is, a specific size or version of a given product — might react to a particular inducement (say, a pricing change) in one region for a single retail chain in a controlled store test under a specific set of conditions. But it could not work through the complexities of analyzing all its SKUs across the entire chain, through all the possible marketing-mix variations. Even with scanner data providing molecular insights into buying behavior, traditional statistical techniques haven’t been able to separate fool’s gold from the real thing; they cannot handle the mountains of data or adequately factor out anomalous events (the weather, traffic tie-ups, etc.), and they consistently generate flawed estimates. Creating thousands of individual statistical models to analyze the combinations of factors at the level of each event would be an overwhelmingly expensive and fruitless task; each sales account rarely has enough data to support such analysis across all the relevant variables.
Incapable of making fact-based budgeting trade-off decisions, marketers routinely have fallen into the pattern of developing and funding their strategy on the basis of history and anecdote; or by using undemanding audience segmentation schemes; or with simple input–output analyses, in which they impute causality to correlatives, assuming a relationship, for example, between awareness and sales. Their analyses — or models, if they build them — will cover entire retail channels, large swaths of geography, or expansive event categories.
Such strategies are inefficiently broad. Even if they are right on average, they will be wrong almost everywhere. An approach that boosts preference or volume in one region or medium or customer group might well depress it in others. A food manufacturer will rightly assume, for example, that a hot-sauce promotion will do better in Texas than in New Hampshire. Closer inspection, though, may show that a Texaswide promotion works well in some store chains, but terribly in others. Why? Perhaps it can be attributed to differences in retailer demographics, maybe the price point is appropriate for one chain’s buyers and not the other’s, or possibly the product ran head-on into a simultaneous promotion for a competing sauce. Standard modeling techniques and planning processes can’t really say.
Two decades into the era of scanners and cheap computing, however, it is possible to analyze data efficiently and rapidly, which enables marketers to explore cross-effects in enough detail to predict event outcomes — and thus improve events — with confidence. Simple regression models that look solely at, say, price/volume relationships can readily be replaced by mixed models, which are able to explore dozens of variables and generate predictions about the impact a specific event will have on a given SKU at a particular store.
These mixed models draw from advances made in such diverse fields as signal processing and approximation theory. In modeling trade promotions, for example, a relatively new statistical estimation technique, Bayesian shrinkage, enables researchers to combine information from numerous retail accounts and SKUs, and provide reliable models without a major computational burden on each individual account and SKU. Bayesian shrinkage has been used across the public and private sectors, to forecast international growth rates, for example, and predict food-stamp program eligibility.
Like the electron microscope, such modeling techniques, although complex, allow atomic viewing in detail: Which retail accounts will respond better to promotions for Product A, but to media advertising for Product B, and so on. With the frontier of knowledge pushed forward, companies today can answer some of the most basic, yet previously elusive, questions they have faced. (See Exhibit 1.)
Advanced modeling techniques can substantially alter the way companies perceive their customer base, and even how they measure customer lifetime value. One media company with which we worked had long used demographic segmentation to predict subscriber renewal rates. We employed a statistical method called latent class modeling, which uses a regression model that includes a hidden class variable, to identify underlying subpopulations that might be more differentiated and predictable in their behavior. From some 200 behavioral variables at the start — which included the channels through which customers initially became subscribers, renewal cycle history, and subscription-postponement history — we were able to define 15 segments. Their renewal rates varied to an astonishing degree, from 0.2 percent to 86.4 percent. None of the high-return segments included demographic variables.
Even in indirect market environments, where demographics are all a company has, effectiveness can be dramatically improved through the analysis of results at a much more appropriate (e.g., geographically targeted and execution-specific) level. Spectra Marketing Systems Inc. uses mixed modeling techniques to pinpoint the impact and ROI of advertising in spurring demand for specific products among specific consumer groups in specific MSAs. By mapping the entire U.S. retail landscape and overlaying it with census-derived demographic information, the company is able to isolate store-by-store effects of television ad campaigns, and judge whether a demographically aimed effort is actually influencing behavior among the target group. The technique is increasingly important as television fragmentation and niche cable allow for more refined planning, and holds the promise of slowing, or even reversing, the long-term trend of marketing dollars’ flowing from advertising to brand-eroding promotions.
Decision-Support Systems
In an ideal world, appropriate analytics would be all that is necessary to solve John Wanamaker’s dilemma. Unfortunately, in marketing, knowledge alone is not power.
Facts abound in the marketing organization. Correct, valid, and useful facts are offered up by advertising giants such as BBDO and J. Walter Thompson; syndicated research firms such as A.C. Nielsen and Information Resources Inc.; and vertically focused research companies such as automotive authority J.D. Power and Associates. But marketers have been unable to take sustenance from this sea of data because the internal and external sources don’t agree on the basic dimensionality and measures of what they record — even if an analytical engine is able to refocus the organization on the measures that matter.
The marketing company’s legacy systems must shoulder a lot of the blame. A typical firm is filled with systems for planning, tracking shipments, tracking customer payments, and so on. Conventionally, each is in a loop so closed that the system speaks only to its departmental acolytes and itself, becoming, in effect, an information bunker in which marketing or sales or finance can hunker down to hurl salvos at its organizational enemies. Siloed systems contribute to pushing sales and marketing away from customer-facing activities; too much time is spent gathering data, generating reports, and pursuing “rates and dates” transactional selling, and not enough time is devoted to relationship building, collaborative business planning, and solutions selling.
Unfortunately, the conventional response to isolated legacy systems — an integrated CRM system — has been equally bad. CRM software typically lacks analytical rigor; it tends instead to capture and embed standardized processes (euphemistically called best practices) in such fields as sales pipeline management or marketing campaign management, and to force companies to follow those processes, with little room for adaptation.
To match knowledge with people and get real results, no matter what the corporate boundaries, requires a precisely defined information system distinguished by three characteristics:
- The tools are very rich in intellectual-property content, tailored to metrics important to the business, and refreshed through time as knowledge develops and needs change.
- The system integrates three components — planning, transactions, and post-event tracking and analysis.
- The system has a simple interface and visibly supports the change program goals, such as the shift from volume at any cost to profitable volume.
Designing a suitable integrated decision-support tool does not require a company to abandon its legacy systems; indeed, it is often better not to take on the cost of replacing existing technology, but rather to add missing capabilities tactically to the present structure. Platform also does not matter. Good systems can be built on an Internet platform, but they can also be constructed on old mainframes, as sophisticated yet frugal companies such as the Capital One Financial Corporation, one of the U.S.’s top 10 credit card issuers, have shown. The benefit comes from the ability of all relevant personnel, from sales reps in the field to media planners in headquarters, to compare and contrast formerly isolated information; system design is an IT strategy issue executives should manage and justify separately.
For example, after designing a product-line management tool that allowed executives at the headquarters of a home furnishings manufacturer to better forecast product life-cycle discontinuities, predict consumer response to design features, and calculate cannibalization factors in the product mix, we developed a sister version for the sales force. This enabled the personnel charged with managing superstore accounts to recommend product substitutions to store managers, which was crucial to allowing this manufacturer to maintain its share of shelf space. (See “Focus: A Case of Category Management,” at the end of this article.)
Redesigned Business Processes
Robust analytics, shared across the various divisions and units through well-designed decision-support systems, allow companies to establish the kind of consistent and aligned processes that circumvent organizational sandbagging and drive profitability.
In most companies, marketing and sales best practices are determined by the least common denominator. Much like the world of Kurt Vonnegut’s Harrison Bergeron, all accounts, geographies, and marketing vehicles are weighted down by budgeting and planning processes that force them to aim for the same chance performance levels, regardless of their intrinsic capabilities.
Consider a typical example, which is abridged only slightly from our actual — and repeated — experience with packaged-goods manufacturers. A food manufacturer has developed a new soft drink flavor. At its annual sales and marketing planning meetings, it sets a national advertising and promotion budget to support the product. It then executes local media buys consistent with the product rollout strategy. Those budgets and campaigns will be spread generally evenly to accounts across the nation, or at best within a region.
Established that way, the targets, almost by definition, cannot be locally relevant. Efficient and effective spending has to account, for example, for the incursion of Wal-Mart Supercenters in a specific MSA, and the consequent decline of, say, Retail Chain X; promotion targets and supporting budgets set regionally will inevitably be misaligned. The manufacturer’s Wal-Mart account rep will have no difficulty reaching her target, so she has little incentive to attempt to negotiate better pricing and higher profitability from her events. Chain X’s rep, however, will view his volume targets in isolation; failing to calculate Wal-Mart’s impact on his account, he will probably spend his entire budget, yet fall short of his goals. Another possibility: He’ll be well aware of the Wal-Mart effect, know that his targets are unrealistic, and just do what he’s told, spending his budget on a self-fulfilling failure.
Before an ROI marketing capability can be instituted in the firm, a marketer has to streamline its processes. The four areas that need particular attention are business strategy and budgeting, target setting, tactical planning, and post-event analysis and tracking.
In business planning and budgeting for trade promotions, a discipline that accounts for a heartbreaking amount of waste, companies can use the redesigned analytics and systems to finally establish accurate spending, profit, and volume targets both nationally and for individual accounts. Target-planning processes then can build the most effective means for reaching these targets — by account, by market, and even by SKU. The plans can make explicit the number and cost of inefficiencies at the account level, with the aim of eliminating them as rapidly as possible. To that end, plans also can be reconciled with — and modified in accord with — retailer objectives. Equally significant, promotional and brand marketing campaigns can be made mutually reinforcing.
Finally, tracking processes must be set up to monitor implementation of and detect deviations from the plans. If results differ from the plan, the sources of variance (was it the snow, display execution, or an unanticipated sale at a competing grocery chain?) can be located. This post-event tracking and analysis can further heal the age-old breach between sales and marketing by spotlighting the successes and failures attributable variously to field efforts, brand advertising, and even product characteristics.
Although the above description of ROI marketing process coordination may sound a little too idealistic, some of the world’s most sophisticated consumer-marketing companies have, in fact, moved aggressively in this direction, by starting modestly, espousing a clear vision, picking discrete targets, measuring the results, and celebrating the successes, thereby creating greater impetus for the next step forward.
One packaged-goods manufacturer, the world’s leader in its primary category, began to repair its process by bringing together its brand managers and its sales executives to plan year-ahead product strategies in tandem for the first time. When that went well, the company opted to take the next, and significantly more difficult, step: realigning its business-planning cycle with those of its retail accounts. Its planning process (which coincided with the manufacturer’s fiscal year) came so late in the retailers’ cycles that the manufacturer was unable to take advantage of opportunities retailers routinely build into their annual schedules — “Dollar Days,” “Presidents’ Day Sales,” and the like. Although the company understood the problem, it believed it was hamstrung by its need to report a detailed financial plan to its board of directors. We were able to show executives that the aggregate variability of trade promotion spend was de minimis; the company could align its planning and budgeting cycles with the retailers’ — meshing its product strategies with their plans to open new stores or stock different sizes — and still report a financial plan with confidence.
Other, equally dramatic, changes also came to pass. The company instituted an annual “Brand Summit,” through which customer teams from the manufacturer’s seven largest accounts contributed to the creation of brand marketing strategies. The company also developed “Planning Centers,” where events planners — who conventionally work in isolation, tethered to individual retail accounts — share best practices and optimize their plans across accounts. The cross-pollination has not only introduced successful concepts to places where they had never been tried, it has also encouraged planners to test and measure creative new ideas.
Organizational Alignment
Developing the strategy, realizing it in conjoined marketing and sales plans, and establishing systems and processes to support them are not enough. Reshaping the marketing, sales, and finance organizations — the internal structures, relationships, incentives, and operating practices that determine how the strategy and plans are enacted — is the most crucial element in finding the better half of marketing spend.
In describing the ROI marketing framework, we make a distinction between corporate processes and organizational alignment because the former largely concern formal group activity, whereas the latter focuses on shaping, bounding, and encouraging individual actions in support of the firm’s objectives. Organizational alignment aims to decentralize tactical decision making, define decision boundaries, and ensure that metrics and incentives gauge and reward actual, rather than apparent, success. Ultimately, proper alignment makes brand and finance managers customer-centric and turns salespeople into business managers who are less directed toward tactical selling and increasingly responsible for account profitability. In consumer packaged-goods companies, this “reskilling” and empowering of the different personnel greatly strengthens the entire enterprise. Salespeople become more capable of dealing with increasingly powerful retailers and better able to contribute to core decisions about product and corporate strategy. Marketers learn to help both emerging and declining channels compete. Finance executives can fine-tune the trade-off between volume and profit.
Consider the case of one client of ours, a media distributor. The company had decided to refocus on its most profitable business, the music category, for one retail segment, major mass merchants. Even with the new strategic focus, the range of products was vast, as were the complexity of decisions and the speed with which they had to be made: A rack jobber, the company worked with more than 5,000 retail outlets, merchandising and managing about 3,000 items in each, out of a universe of nearly 300,000 compact disc and cassette titles. Profitability depended on keeping the right products in stock, tailoring selection to consumer demand by retail account and by individual store, and lowering the product return rate.
As originally designed, the organization’s structure installed powerful customer teams charged with satisfying the mass-merchant customers. Faced with making so many individual decisions for such a large number of stores, the new customer teams had to rely primarily on national average data. These teams also coexisted with a strong central purchasing department, which largely controlled which titles were bought and in what quantities. Given its established relationships with the music companies, purchasing maintained its ability to “force out” product in response to breaking hits.
The field personnel, who visited individual outlets, were reduced to functioning as an execution arm of these two centralized, competing decision makers. Opportunities to stock titles or launch promotions that might be effective in some subset of stores were often overlooked. At the same time, other stores consistently received too much product in specific music genres because it was impossible for central decision makers to customize selection at the individual store level.
As part of a comprehensive ROI marketing program, we recommended the creation of a core merchandise-planning function. Its sole purpose is to coordinate transactional decisions between purchasing, the customer teams, and the field. Merchandise planning both responds to field-generated requests and decides when to order merchandise for a specific store on the basis of information that may not be available to the individual sales rep. This function provides a central coordination point for all order streams and, aided by decision-support systems that optimize title selection in the stores (by clustering stores and building a model that indicates where to add, pull, or switch titles), it can spot and react to trends in demand patterns.
In addition, to develop customized responses for individual situations, we designed and implemented a new organizational model featuring empowered district managers, each of whom runs, in effect, a 60-store music chain. They have input into nearly all decisions that affect product flow into their stores and are encouraged to develop new and innovative ideas, such as niche product merchandising. (See Exhibit 2.)
This new organizational model has had a powerful effect on the distributor’s business. In the first year following implementation, the company achieved revenue growth of 10 percent in an industry that grew by just 2 percent. Operating profits have nearly tripled.
The Big Idea
Addicted to a reigning ideology — that “persuasion is not a science, but an art,” as the renowned advertising executive Bill Bernbach once put it — marketing executives, almost from the beginning of mass marketing, have believed that they should intuit the “Big Idea,” and all else would follow. In fact, though, they have been trapped in a cycle of assumption, approximation, and acceptance.
They assume that one answer works everywhere — because surely tailored strategies can’t possibly pay off against the complexity involved in designing and implementing them.
They approximate the return on their massive selling and marketing expenses — because one can never know with certainty which half works.
They accept that their marketing and sales organizations will be in constant opposition because of this inherent imprecision — and that the job of the business leader is to force the brutes to stay on strategy.
Worst of all, they miss the Big Idea when it clubs them in the nose. Without precision and understanding, all ideas are simultaneously big and small — big if they work, and small if they don’t.
ROI marketing allows the facts to separate the giants from the midgets — quickly, precisely, and clearly. It affirms that the barricades between information silos can come down; that everyone in the organization can share common data and analytical tools; and that marketing and sales strategies can be transparent, measured, and adapted on the basis of real results, not historical anecdotes.
It is happening in a growing number of companies, in the U.S. and abroad. By laying a foundation of fact inside the firm, ROI marketing allows CEOs and business unit managers to concentrate less on calming the troops and more on relating to customers; it helps brand marketing executives appreciate the thrill of the successful sell; it enables salespeople to grow into business managers; it allows the finance team to participate more in the strategic development and prosecution of the business.
As marketing’s ancien régime continues to crumble — as television audiences fragment, ad recall scores decline, and brand equity erodes — ROI marketing will become a requirement for survival. The better half will prevail. And at long last, old John Wanamaker will be able to rest in peace.
Reprint No. 03104
Focus: A Case of Category Management |
Since the 1980s, the U.S. building-products industry has been rocked by the emergence of “big box” superstores. Like “category killers” in other retail sectors, home centers like Lowe’s and Home Depot have flourished by offering one-stop shopping, lower prices, and gratification more immediate than could be provided by smaller specialized retailers. One manufacturer of a core interior product found its position compromised by the increasing importance of these home centers. Although the manufacturer’s scale, product-line breadth, and brand name gave it some protection, copycat competitors with lower costs and the superstores’ negotiating power conspired to drive down its margins. Moreover, although the superstores carried actual inventory (whereas the smaller retailers generally offered only order books), the selection available in the giants was only a fraction of what our client produced. The best — perhaps the only — response the company could make was to give the retailers the capability to command higher sales per unit of shelf space, thus persuading them to accept lower unit margins in return for greater total sales. To drive higher volume through the same shelf space without dramatically lowering prices meant the company had to place the right mix of items on the shelf for the people who shop at specific store locations — no easy task in a business that offers literally hundreds of design combinations at a time, with fully one-third of the SKUs replaced annually. Historically, the company’s product design decisions were based on designer intuition. New designs were tested almost entirely by qualitative means; even though some 80 percent of the resulting new products failed, the traditional instinct-test-pray approach was thought to be the only one possible. The company did no modeling, but instead used static sales reports to judge product effectiveness from period to period. Assessment of trade-offs — what might new Product P do to existing Product Q? — was not possible. As part of an ROI marketing program, the manufacturer conducted extensive market research on how consumers were attracted to specific attributes (color, pattern, price preferences, etc.). The research revealed that by far the first and second priorities of consumers were color and pattern, in that order, and that neither price nor store location made much difference. We embedded these and other customer understandings (e.g., consumers were far more willing to substitute black for white, and vice versa, than make any other color substitution) in an algorithm that meshed product-choice decisions with the customer’s priorities. This analytic engine was able to predict consumer responses to attribute-mix variations in different demographic areas. Next, the client built a decision-support system that allowed the company and its designers to optimize their design decisions. Among other things, the methodology, borrowing from quality-control theory, could identify products coming to the end of their life cycles, which allowed the company to become more aggressive in substituting products before they passed their tipping point into the rapid-decline phase. Because it had this data before the retailer saw it in sales figures, the manufacturer could offer its own products as substitutions for the declining designs. We also created a product-line management tool, which used the same central data warehouse, that allowed sales reps to match product with a specific store’s consumer demographics. Ultimately, the system could forecast how volume and profitability for both the manufacturer and the retailer would change in response to a proposed shift in the product lineup; which products would be cannibalized by the new addition and which ones should be removed. All this could be done by the rep in the store, on a laptop with the retail buyer. Category management, which is usually based on “fair share” arguments — i.e., if my product makes up 20 percent of your sales, then I should have 20 percent of the facings — was transformed into a business discussion. Sales calls that had traditionally focused on case pricing became centered on gross margin, inventory turn expectations, and consumer preferences. — L.H.M. and S.K.M. |
Authors
Leslie H. Moeller, moeller_leslie@bah.com Sharat K. Mathur, mathur_sharat@bah.com Sharat K. Mathur is a principal in Booz Allen Hamilton’s Chicago office. His areas of expertise are customer and business strategy, pricing and trade promotions, marketing effectiveness, and helping clients develop customer-focused organizations. Randall Rothenberg, rothenberg_randall@strategy-business.com |