Segments in Time
The trouble with most market segmentation programs is that they capture customers' preferences only at one point in time. Instead of snapshots, what is needed are moving pictures -- or what the authors call "evolutionary segmentation." By detecting unfolding trends, companies can gain a much better understanding of what customer preferences are likely to be tomorrow, and why.(originally published by Booz & Company)
Businesses and markets once remained the same for decades. Now change is rampant. To succeed, companies must understand the logical foundations of market and business evolutions.
Not long ago, a large telecommunications company wanted to know more about the buying habits of its long-distance business customers. The company hired a consultant to divide the customers into discrete market segments. In due course, the consultant produced an impressive study analyzing the customers' needs on 15 fronts and identifying almost 20 "distinct" segments.
Barely six months later, however, the long-distance market had shifted, causing customer needs to change and leaving all those neat categories behind. The company had to bring the consultant back for an update, and keep bringing him back every time the market changed -- which it was doing with increasing frequency as the pace of technological innovation accelerated. What's more, for this segmentation approach to be used effectively as a marketing tool, every prospective new customer had to be exhaustively surveyed about where it fit in terms of those 15 different needs. And that was a task that the company's overburdened sales force undertook only reluctantly, and sometimes not at all.
Frustrated, the company hired another consultant, who developed a different segmentation system -- this one focusing on customer behavior, rather than needs. Relying on a simple four-box matrix, the new approach seemed to promise the company's managers the means to quickly decide which customers were good targets and which were not. (See Exhibit I.)
Typical Static Segmentation: Long-Distance Corporate Customers
Source: Booz-Allen & Hamilton
Alas, the company discarded this scheme nearly as fast as the first one. The new segmentation proved no better at helping managers keep up with a constantly shifting landscape, since it offered little insight into why some accounts were vulnerable, how customers were changing or what was necessary to beat back the competition and hold on to market share. Thus, the company continues to struggle to better define its customers and defend against competitive inroads.
The Right Kind of Segmentation
This experience is not unique. Many companies have adopted segmentation schemes, often developed by the best and the brightest, only to soon discard them. Why? Perhaps because almost all of these analyses are static. They provide a snapshot of the market -- of customers' preferences at the moment. Yet what companies really need is a moving picture, an understanding of what those preferences are likely to be tomorrow, and why.
Market segments evolve over time. They may grow, decline or disappear. They change for a number of reasons: new demographic trends, technological advances, increasing competition, regulatory impositions and economic cycles. And, of course, the individual consumers and companies that make up the segments are themselves changing, leading to shifts in their needs, purchasing patterns and loyalties.(1)
Thus, rather than analyzing customer segments that existed in the past, or at one point in time, marketers, planners, sales managers and senior management should think in terms of "evolution." Segmentation ought to incorporate changes in the market, especially when the pace of change is rapid. Further, a good segmentation should help explain and predict the reaction of both companies and customers to changes in the marketplace environment. Understanding these patterns is likely to become an even more important tool for successful managers in the future, since companies will prosper only if they change along with their markets. (See Exhibit II.)
Comparison of Segment Understanding
Source: Booz-Allen & Hamilton
Through 700 interviews at more than 200 companies in industries as diverse as computing, petrochemicals, retailing, telecommunications and transportation, Booz-Allen & Hamilton confirmed that evolutionary segmentation can be a powerful means of improved performance. This article reports three major findings, all of which suggest a focus on tomorrow's customer segments, rather than yesterday's:
The time when a customer is moving from one segment to another is particularly important.
Change comes in steps, rather than evenly and gradually over time.
Evolutionary changes are predictable for groups of customers, but predictions require careful consideration of the internal economics, objectives and politics of customers and potential customers.
How Segments Evolve New Segments
Corporate leaders must look forward: What new customer segment is being created? How do you address this new segment's needs? If customers are constantly changing, how can you retain them? The most successful companies have learned how to adapt their market approach to take advantage of market change.
Consider one market as an example. Chemical companies such as Amoco, Tenneco Packaging, Formosa and others sell wraps, food-service utensils and Styrofoam containers to wholesale distributors such as Bunzl, Kraft, Sysco, Northeastern Bag, Marstan and thousands of smaller distributors. Distributors, in turn, sell to schools, institutions, restaurants, office buildings and smaller supermarkets. Initially, most distributors were small or medium-sized family-owned enterprises, with family members holding key positions and younger relatives positioned as heirs-apparent. However, the economic driver in this business is cost, and larger entities can purchase supplies more cheaply from manufacturers through volume discounts. This leaves small independent wholesalers with four choices:
combine with other distributors to become a "super independent."
sell out to national operators like Bunzl or Sysco.
join a buying group, with different levels of autonomy, like Resourcenet or Unisource.
remain independent and hope for the best.
Each of these options represents a customer segment for vendors like Amoco or Tenneco -- and which option a distributor chooses is key to the marketing approach that is needed. A national wholesaler, such as Bunzl, is more price-oriented and centralized in its purchasing choices. Independents, by contrast, stress service, credit and relationships with manufacturers. Thus, a vendor faces a very different selling environment and set of price pressures depending on the distributor.
The timing and likelihood of a distributor moving from one segment (e.g., independent) to another (e.g., part of a national company) depend on a number of factors, many of which are observable to vendors. For instance, competitive pressure led one distributor to sell out recently in the Philadelphia market. But because of the desire to hand over the business to next-generation family members, another distributor there chose instead to remain independent and joined a buying group. Note how this choice is dichotomous: once in a buying group, a distributor is less likely to sell out because it can now buy at lower costs. Similarly, once a distributor is part of a larger company, there is no need to join a buying group. (See Exhibit III.)
Evolutionary Segmentation of Plastics Distributors
Source: Booz-Allen & Hamilton
Manufacturers can leverage an understanding of the evolution of this business in a number of ways. Those selling to independent distributors can look for signals that their customers may jump to new segment categories, e.g., family members drifting out of the business may signal a potential buyout. In that case, the next step is for the vendor to cultivate a relationship with the company's future decision-makers at a national distributor's headquarters, or elsewhere. The key point is that the time to win or lose an account is before and during the time a distributor makes a segment change -- this is when contracts and buying procedures are up for grabs.
Vendors to markets undergoing change need to try to anticipate their customers' evolutionary drivers and consequent actions. This requires an understanding of the mind-sets of consumers, or of what is going on within corporate customers. The data processing/data communications field provides an important and illuminating case study of the value of understanding how customers change, and how segments evolve over time.
A Case Study with Five Steps: Data Communications Equipment
The orientation of business data users is evolving from mainframes to personal computers. This evolution is not gradual, however. There are five distinct phases that companies can pass through, each of which represents a segment for data communications vendors. Some users may pass through a few stages, while others go through all five.
The first phase, which everyone began with a few years ago, is a mainframe-only approach, featuring a central Management Information Systems (MIS) department that controls all purchasing and deployment of computers. This is the Hierarchical Proprietary stage.
Many companies next experience a radical policy and power shift in information management: the end users of the computers decide to buy PC's, minicomputers and other equipment, bypassing the MIS department. Inevitably, after buying the computers, the end users try to connect (or "network") these devices, so they begin to lease lines and set up networks that are also independent of the MIS department. For instance, at one bank, United States Trust, the MIS department conducted a survey a few years ago and was surprised to discover that more than 230 networks had been set up by end-user departments. This phase is called Mixed Architecture.
The next phase is an organizational response to the chaos produced by such mixing and matching. Companies form an independent central telecommunications department, or utility, that tries to regain financial control and reduce costs. It does that through the consolidation of purchasing and the physical consolidation of networks. This is called the Independent Telecom Department phase.
In the fourth stage, coming after a period of perhaps several years, companies integrate telecommunications and computing again within the MIS department. The new arrangement typically allows end users greater access to information, which is now kept in large corporate data bases administered by MIS. To access these large data bases, high-capacity data networks are created to allow users in remote locations to pull down large files through a corporate "Internet" and work on them on-site(2)
For instance, the pharmaceutical manufacturer Merck & Company allows "mission critical" staff, like researchers, to pull huge molecular models out of data repositories, downloading the material to their work stations. This phase is called Mission Internet.
The last stage is either an alternative or a successor to the fourth stage. Instead of giving over the corporate data bases to the MIS department to place in a few large repositories oriented to the most important end users, the company adopts an architecture to allow all end users to access any data they need. This means throwing out a lot of obsolete equipment, and introducing a homogeneous infrastructure in which all the components can interact, or "talk," with one another. This is technologically quite complex.(3)
To pay for the new equipment, and endorse the technological risk of this architecture, non-MIS business leaders must be intimately involved and convinced of the need for the expenditure. Thus, the new power structure is a close partnership between MIS and functional departments. This stage sees departmental computer specialists stripped of some of their autonomy as a cohesive new approach to purchasing and technology takes hold. This final stage is called Integrated Computing. (See Exhibit IV for an overview of all five stages.)
Data Communications User Segments
Source: Booz-Allen & Hamilton
Evolutionary Logic Chain
The initial shift from the monolithic architecture of the first stage sets in motion forces that inevitably propel a company up the evolutionary chain, link by link.
The predominantly technical orientation and multiplicity of networks in the second stage is expensive -- sometimes accounting for more than 10 percent of a company's total cost base. This provokes management to reduce costs. Thus, many companies cede control back to MIS or to a newly created independent telecom department. Which route they choose depends on whether management can agree on a common architecture. It also depends on whether MIS has managed to develop the flexibility and responsiveness that initially caused end users to take on the job of buying the equipment themselves.
But because the new telecommunications department often has a strong focus on minimizing costs, there is usually less focus on innovation. This resistance to new technologies, many of which would benefit line operations, causes line managers to look for alternatives to the "penny pinching'' telecom people. This pushes the company to move to the next evolutionary change, in which dissatisfied end users are wooed by a new, more enlightened MIS department. Business managers looking for new capabilities, and MIS managers looking for greater architectural consistency, drive department heads to integrate the telecommunications/data processing infrastructure and make it more forward-looking. This pushes them into phases four or five.
The choice of whether to move to Mission Internet or Integrated Computing depends on the company's business requirements and opportunities.
If its overall return on investment is high, then the company will most likely choose elegant and uniform data base architectures. This was the case at Morgan Stanley, the investment bank, and at Federal Paperboard, the specialty manufacturer, which both had ample cash to invest in internal information technology. Typically, this architecture is administered by MIS but oversight is by MIS and senior line managers, vice president and above, who are intimately involved in funding choices and alternatives.
Conversely, many larger enterprises with a lower return on investment cannot afford to replace the existing hodgepodge with a clean new data architecture, which usually requires all new software and often a major makeover in hardware. In that case, line managers hand over the problem to MIS. That is what happened at the Chrysler Corporation and other large manufacturers.
The strong internal logic of the evolution drives buying behavior. In the move from the Hierarchical Proprietary stage, departmental computer specialists seek immediate solutions to interconnection requirements. Thus the performance of equipment and ease of use are key. In the move to the Independent Telecom Department phase, cost and centralized control features play a more important role. In the shift to Integrated Computing, the focus is on the ability of equipment and vendors to perform smoothly in conjunction with software packages and to innovate to meet end-user needs.
To highlight how different these requirements are, note that a low-cost bidder is likely to win only in selling to the centralized telecom department. Thus, it would be foolish to deeply discount pricing in bidding in the other two environments. Vendors must understand this logic (low cost, high capabilities, etc.) in order to preserve margins and service different segments. This requires an overall road map of segments and potential migration paths. (See Exhibit V.)
Data Communications Evolution
* TCP/IP is a data architecture defined by the International Standards Organization, which is oriented to distributed data architectures such as personal computers. It is the standard used on the Internet.
Source: Booz-Allen & Hamilton
Why Isn't Evolutionary Segmentation Used More Often?
Although most industries appear to evolve in a coherent fashion as managers adapt to the same technological, social, financial and regulatory changes, corporate strategists rarely exploit this fact. Planners routinely identify "key industry trends" on an annual basis, but do not always tie these trends to business plans and targets when analyzing market opportunities. For analysis, planners often revert to more limited but safe and convenient segmentation categories such as Standard Industry Classification (SIC) codes or customer size.(4)
Edward Teller, the co-inventor of the hydrogen bomb, once remarked that the reason the bomb had not been invented sooner was that no one was aware that it could be made.(5) Similarly, a major reason most managers do not employ an understanding of industry evolution is that they are not aware it can be done.
Another reason is that many marketers do not have the time or resources to invest in the technique. Evolutionary segmentation does require a little more upfront investment of time and money to capture and digest trends. Consequently, planners make do with less resource-intensive -- but shorter-lived -- methods.
What Evolutionary Segmentation Is Not
Evolutionary segmentation involves more than simply purchasing prepackaged groups of sociological/psychographic/demographic classes. These prepackaged collections will rarely bear any resemblance to groupings formed by fundamental forces changing a company's market.(6) Superior segmentation is based on understanding the underlying reasons for the creation of new forms of demand.
Evolutionary segmentation is analogous to life-cycle marketing, but different, too. Life-cycle marketing is used by insurance companies and others to identify prospects.(7) For instance, insurers track graduations, marriages and births because these events often signal the need for life or health insurance.
Evolutionary segmentation is different because industries do not ever repeat a "cycle" in that predictable sense. Market segmentation must change with broad industrial and social changes, which have the effect of constantly shuffling the deck for all the players at the table. For instance, software companies today are subject to greater economic, competitive and quality pressures than they were 10 years ago -- forcing them to respond in different ways as well.
How to Do It
Evolutionary segmentation, though a new approach, is not difficult to implement. To begin the process, senior management must insure that planners look for the logic shaping the evolution of their (or their clients') particular industry.
This means that management must choose the right scope of inquiry, as described below. Planners and managers must then "get their hands dirty" by conducting many industry interviews. These will be helpful in establishing some patterns. For example, those interviewed might say, "We used to be centralized, and tried to approach the market through a highly disciplined process, but then we met a lot of resistance and the department was split up among several other departments and everyone does it their own way now." Or they might say the opposite, or some variation in between. The point is to conduct enough interviews to have a large inventory of case histories.
The next step is to make sense of these case histories. Why a company originally tried one approach and switched to another requires detective work. What was the internal political process? What was the operational advantage? Did this company note a similar pattern at another company and copy it? What were the triggers to the migration from one segment to another?
The goal is for planners and managers to become familiar with how finances, operations, strategy and organization all interconnect. (See Exhibit VI.) If they are to predict change, they need to understand all the reasons why a specific organizational shift, for example, must precede an operational one. This detective work can be assisted by financial modeling, by organizational behavior theory and by an understanding of macroeconomic and societal changes affecting the industry.
Source: Booz-Allen & Hamilton
A simple test of whether the process worked is whether planners are able to articulate the richness of likely evolutionary paths -- that they are able to show what factors led specific organizations to specific technological or operational approaches.
The final step is to develop a road map of how companies have progressed over time. In the case of data communications customers, for instance, the migration steps are complex. Map the most common steps, and pay particular attention to companies that have undergone several migrations and tend to be leading edge.
Unlike static segmentation, which offers little structure in quantifying future market proportions(8), evolutionary segmentation allows accounting-like accuracy over time. That is because "inputs" (earlier stages of evolution) plus growth must equal "outputs" (later stages of evolution). In other words, evolutionary segmentation tracks where specific customers come from and go to, and so gives a better indication of what segment growth shares will be like.
How to Segment Consumers
Like segmentation of companies, segmentation of consumers begins with the observation that something is changing in buying patterns. Some changes in the market are insubstantial (long or short hemlines), while others are material (self-service gasoline stations encroaching on full service).
How do you know if the change is fundamental? Evolution in consumer markets is more difficult to discern than in industrial markets. Managers must look at the range of customer factors: economics, product/service needs, customer sophistication, fundamental social change, etc. These will suggest what the drivers of change are and begin to explain why change is occurring.
In some cases, leading-edge consumers are followed by their mainstream peers. For instance, sophisticated users of the Internet have led the way for others to obtain music, telephony and product information over this medium. In other cases, one generation of consumers retains its preferences and behavior, but succeeding generations shift to new patterns of behavior. For instance, younger shoppers are often more comfortable with self service than are older shoppers. In either case, harbingers of change should be heeded, for they signal the next stage of evolution for a particular individual or group.
Knowledge Is Required
More information is required to initially develop an evolutionary segmentation than a static segmentation. The latter can often be based on internal sales figures and commercially available statistics. Evolutionary segmentation requires more outwardly focused data: for example, a view of the entire market is necessary to account for segment growth and customer migration. Also, cold statistics will rarely suggest the right ("natural") segments.(9) Thus, as noted, extensive interviews are required -- of both one's own customers and those of competitors. We find that focus groups often help in this process.
Routinizing the process requires that analysts track the histories of customers over time. That means current customers should be categorized by their current segment status. In most industries, the sales force should assist in the classification process. Senior management should not allow this to become merely another unrewarding task done to humor the marketing staff: sales management should use the classifications to allocate resources and set targets.
To capture fully the benefit of an evolutionary segmentation approach, sales management should dynamically adjust its account strategies according to potential sales opportunities. One data communications manufacturer adopted an approach of scouting for signals of an impending segment migration at potential customers. If the manufacturer found such signs of impending migration, then it dramatically increased account resources -- aimed at likely future decision-makers.
Successful institutionalization of evolutionary segmentation will generate a unique store of knowledge of customer trends that won't be easily duplicated by competitors. In contrast to SIC codes, evolutionary segment data bases can be redefined and refined over time. The company can observe new segments and evaluate the customer base for potential migration to these segments. In a few years, the company will have a unique, proprietary insight into the history, transformation and future of the market while competitors will have to constantly start anew with yet another snapshot to be discarded just as the market changes again.
Benefits of Segmentation
There are three benefits: reaching attractive customers, understanding how to appeal to them and designing new products and services for those segments. Most important, the process helps companies recognize opportunities before competitors do.
Being slightly ahead of the competition is a huge advantage.(10) By utilizing evolutionary segmentation, a company can leverage its understanding of the marketplace to deploy products and services in the right way at the right time. Examples of "early is best" include:
Best Buy. Originally an also-ran electronics superstore, Best Buy boldly innovated to address the increasingly sophisticated segment of customers who desire a friendlier, less-pressured purchasing environment. By being the first among specialty retailers to target this segment, Best Buy outperformed its traditional competitors. Best Buy's compound sales growth rate (1991-1994) of more than 75 percent per year outstripped that of industry leader Circuit City. And it contrasts dramatically with one-time peer Silo-Fretter, which recently filed for Chapter 11.(11)
MCI Communications. In its search for high-volume customers, long-distance carrier MCI was the first to focus on Asian immigrant groups. As these groups grew in number and prosperity, MCI became the leader in such highly profitable segments as the Asia/Pacific calling market.(12)
A leading computer equipment manufacturer, which reorganized its entire sales, marketing and product development organizations around evolutionary segments. It targeted two growing and receptive segments and thereby halted market share losses despite an aging product line.
Tenneco. The chemicals and plastics division of Tenneco used segmentation to identify customers willing to pay a premium for its high-quality and higher-priced products without requesting custom features. Evolutionary segmentation showed that such a target group, while shrinking in number, would still be a viable market over the next 5 to 10 years. After sales priorities were changed to better target this segment, the division's overall profitability jumped in the first year.
Most existing segmentation analysis is reactive, and tries to optimize where a product should be positioned or where the sales force should call, given the status quo. As has been shown, evolutionary segmentation uses the status quo as a departure point and thus can prove far more useful in sales management, product development and marketing functions.
An understanding of evolutionary paths can turn to gold for a vendor's sales force, for instance. That is because every step taken on the evolutionary journey changes what matters inside a potential customer's organization: decision-makers, buying criteria and budgets. The sales force that can anticipate these changes can get to a new decision-maker before competitors do. Arriving first, they can begin shaping the decision-maker's time frames and criteria.(13) When the laggards finally arrive at the new decision-maker's door, they may find that it has already been closed.
Evolutionary segmentation can also produce a big payoff when used to help define a new product or service. Typically, a product cannot meet every potential customer's needs, so some choices and trade-offs must be made. Evolutionary segmentation will help managers determine which segments will grow, and which will be the most attractive sub-markets -- and when.
The drivers of evolution are also the determinants of buyers' preferences. This allows product managers to avoid reliance on that notoriously inaccurate device: customer self-reporting.
When asked, most customers will say they are concerned about cost. This is not always the case, however. As described above in the computer/data communications case study, the values of buyers within the same company vary by the stage of evolution, ranging from long-term investment protection and technical performance to addressing end-user concerns. Only the independent telecom "utility" in the third phase is a short-term cost minimizer.
Superior segmentation can help senior managers identify growing new segments. This can provide an inroad into previously impenetrable markets that are guarded by incumbents who are focused on declining segments.(14) For instance, the start-up Cisco Systems used the growth in personal computers and the ensuing impact on data communications to become the pacesetter in the area of "logical" networking -- leaving formerly dominant data networking players behind. This growth was clearly forecast by Booz-Allen in early 1993, based on the evolutionary segmentation perspective described in this article.
Today, most segmentation is static -- a snapshot in time. While this analysis has utility, it pales when compared with a fluid approach that produces pictures that move.
Armed with a sense of how an industry is evolving and why customers will change over time, management can begin targeting the customers that it wants over the long term. Acquiring the right set of customers is often the fundamental key to improved profitability. Correct segmentation makes the difference between servicing a growing segment with a high willingness to pay for a company's special attributes versus responding to a heterogeneous collection of needs with varying profitability.
Vendors who understand and categorize customers by stages of evolution can reap huge rewards. That includes wooing customers away from established competitors, thus improving sales growth rates.
By rigorously isolating industry trends, observing how customers have reacted to these trends, recognizing what stage a customer is in and predicting potential next steps, vendors may understand customers' future decisions earlier than the customers themselves -- and far sooner than the competition.
Reprint No. 97105
(1) Robert S. Duboff, "Marketing to Maximize Profitability," Journal of Business Strategy, November/December 1992, pp. 10-13. Observes that in today's environment, it is difficult to keep customers loyal.
(2) The means of connecting data bases that allows flexible downloading and uploading of information is called TCP/IP, or transmission control protocol/Internet protocol. The method permits large files to occupy wide network bandwidths, thus speeding the transmission of the files. See Richard H. Baker, "Networking the Enterprise" (McGraw-Hill, 1994), for a review of data networking choices.
(3) The goal of allowing everyone access to the same information presents the problem of keeping the same file up to date with multiple, inconsistent changes. The ideal is a "synchronized" data base joined by a flexible data network. To accomplish this goal, companies like Morgan Stanley are creating flexible relational data bases, using Sybase or Oracle, and improving workers' access to these data repositories.
(4) Benson P. Shapiro, "How to Segment Industrial Markets," Harvard Business Review, May/June 1984. The most difficult, but necessary, part of industrial marketing is assessing the purchasing approaches and the traits and characteristics of companies. This requires creativity.
(5) Speech by Edward Teller, Stanford University, April 1979.
(6) Shapiro, supra. Lists a number of ad hoc segmentation criteria.
(7) See, e.g., Lawrence Lepisto, "A Life Span Perspective of Consumer Behavior," in Advances in Consumer Research, ed. E. Hirschman and M. Holbrook, XII (1985), p. 47ff.
(8) Barbara G. Cohen, "A New Approach to Strategic Forecasting," Journal of Business Strategy, September/October 1988, pp. 38-42. Forecasting is often haphazard
(9) Shapiro, supra. Make decision of usefulness of approach and data availability separately.
(10) Philip Kotler and Paul J. Stonich, "Turbo Marketing Through Time Compression," Journal of Business Strategy, September/October 1991, pp. 24-29. Early is better.
(11) Best Buy's current poor margin performance is a result of some bad geographic expansion choices, in the opinion of analysts. This does not detract from the usable lesson of the company's early segmentation choice and resulting growth.
(12) The tactics were employed by MCI and other marketers in a range of ethnic and niche markets. See Robert G. Docters, Raul Katz and Carolina Junquiera, "Strangers in Their Own Land," Telephony, July 31, 1995, pp. 18-19. See also, "Special Report: Marketing to Hispanics," Advertising Age, Feb. 27, 1986, pp. 11-51.
(13) See Jim Holden, "Power Base Selling" (John Wiley & Sons, 1990). This book describes the importance of identifying and influencing purchasing decision-makers.
(14) See Gary E. Willard and Arun M. Savara, "Patterns of Entry: Pathways to New Markets," California Management Review, Winter 1988; John L. Ward and Stanley F. Stasch, "How Small-Share Firms Can Uncover Winning Strategies," Journal of Business Strategy, September/October 1988, pp. 26-31.