Fish gotta swim. Birds gotta fly. And service companies gotta spread out; geographic expansion has been, and will continue to be, one of their primary means of growth. Adding new retail stores, sales branches, and service centers significantly increases customer access and can do the same for sales. But managing the result—a classic distributed service network with hundreds or even thousands of service and retail-customer touch points—can be surprisingly difficult, and the challenge becomes more complex the more the network grows.

Anyone who has managed such networks will recognize the varied and tough decisions they require. Consider a few typical ones:

• After a rival equipment-repair company announces a two-hour service guarantee, its business jumps. Your company serves its customers in about three hours. Should you do nothing, match your competitor’s offer, or try one-upmanship?
• You are the regional manager of a coffee chain deciding whether to add a tenth store in a given city or to branch out to a totally new one. How do you determine the acceptable level of cannibalization for your existing stores? Do you open the tenth one or enter the new city?
• Last year, some bank branches in your region easily hit their 8 percent growth target, while others worked hard but fell short. This year, you have been told to expand revenues in your region by 10 percent. You think some branches should shoulder more of the load. How do you set differentiated targets?
• As the CFO of a clothing retailer, you have allocated enough capital to open 20 additional stores this quarter. Do you encourage your real-estate development team to select the best potential locations in all of your current markets or to focus on a particular region?

From the cash register to the corporate center, problems like these vex managers, who must balance customer service levels against store margins, determine priorities for capital investment across local markets, and, ultimately, try to wring the greatest profit from geographically dispersed networks. Meanwhile, these managers must also understand that the needs of customers and the nature of the competition vary widely from one local market to another.

In practice, many companies use general rules of thumb or centralized corporate mandates to run their network operations. Lacking quick and easy ways to generate tailored solutions, these companies base decisions about staffing levels, growth targets, and the like on broad, company-wide guidelines, including "one-size-fits-all" expense parameters or financial targets. Thus, for example, a clothing retailer might keep labor outlays at 25 percent of its sales across the board or set uniform sales growth targets of 5 percent a year for all of its stores.

Typically, such mandates fail. Centralized decision making oversimplifies the wide variations among a network’s customers and competitors and isn’t sufficiently flexible to accommodate different growth rates in different locations and regions. In fact, since the network of locations accounts for 50 to 75 percent of the cost structure of a geographically dispersed company, the effect of seemingly minor mistakes in the rule-of-thumb way of allocating resources can quickly be amplified across several thousand locations and perhaps threaten the whole company.

Start at the grass roots

Many managers totally avoid simulation models or queuing theory, believing they offer only marginal insights into strategy

Analytical tools to help optimize a distributed service network have existed for a long time. But companies rightly view them as too inaccessible and abstract to be useful or too time-consuming to use meaningfully at every location. Many regional managers approach customer segmentation, for instance, by relying on the company-wide analysis of the needs of their customers, since they believe that expensive efforts to segment local customers won’t yield actionable or profitable insights. Moreover, many managers totally avoid simulation models or queuing theory, regarding them with trepidation because they are highly technical and time-consuming and believing that, in themselves, they offer only marginal insights into business strategy.

But help lies at hand for beleaguered service network managers. By taking advantage of existing tools that have been combined into a new and more integrated approach—micromarket modeling—they can boost returns at all levels: individual service locations, local and regional markets, and, ultimately, the overall network.

Unlike typical top-down approaches to managing service networks, micromarket modeling examines the business dynamics at the grass roots of organizations—where they meet their customers—and applies the insights gained in this way across the network. The approach takes customer segmentation as its base and adds resource-capacity analysis and financial modeling to determine the financial potential of a service location, the right geographic coverage within a region, and the relative potential of different markets. This combination generates a new set of easy-to-use visual analytical tools—service-performance curves—that can replace intuition and fear of the unknown with a more informed understanding of the dynamics of service businesses (Exhibit 1).

Micromarket modeling is novel for two reasons. First, it creates a set of easy-to-understand visuals drawn from customer, operational, and financial analysis. Second, it gives service network managers previously inaccessible insights and more robust information that justify the time and energy needed to use long-shunned analytical tools. Those insights help managers to unleash location-level economics and capture each location’s economies of scale, to optimize local-market service-model configurations, and to develop informed local-market and network strategies, such as priorities for capital investment not only within and across a company’s existing markets but also in currently unexploited ones.

The case of a national field service company—let’s call it DSN—shows how the micromarket-modeling approach can be used and how building the analysis from the ground up can produce a far more sophisticated understanding of the right way to expand distributed service networks.

Putting micromarket modeling to the test

DSN is a national chain of retail stores that sell appliances and replacement parts to home owners. Over the past few years, several dozen pilot stores experimented with the idea of delivering these products to business customers and specialty repair shops. DSN entered this new business conservatively, expanding it while limiting the company’s financial exposure. Each pilot store received a single van with a driver and one phone representative and was told to maximize sales with existing resources before asking for more of them.

With minimal support and direction from headquarters, the stores started serving a variety of customers that weren’t necessarily the closest, the largest, or the most valuable. The customer base was spread out over a wide geographic area. Customer service goals weren’t differentiated among customer segments, nor did the company have a detailed understanding of its service performance (for instance, average delivery times). Over the course of a few years, stores that happened to exceed the corporate hurdle rate—those in less competitive environments or with more capable staffs—received extra resources, while others, whose sales didn’t grow, were limited to their initial allocation of funds.

Despite gross revenues of more than $300 million, the chainwide returns were disappointing. DSN faced a decision about whether to continue these investments and, if so, at what level. A three-step micromarket-modeling approach helped the company develop a growth strategy contrasting sharply with the strategies generated by its traditional top-down approach to setting targets. DSN is now poised to invest heavily in this new business and has set ambitious growth targets: to quadruple its revenues and triple its earnings margins.

Step 1: Know customers’ breakpoints

The first step in optimizing a distributed service network is to understand the company’s current and potential customers and the threshold of service that dramatically influences their purchasing decisions—their service "breakpoints." A general list of customer preferences (a service bundle of price, service times, and product quality, for example) usually isn’t enough to help companies make informed business decisions, because the expectations of customers range widely across segments and geographies. For an equipment-repair company, a decision about whether to react to a competitor’s offer of faster service depends largely on whether it is really a breakpoint for current and potential customers. In the case of DSN, most of its higher-value customers wanted a "total-solution" service bundle, and the breakpoint was to provide deliveries reliably within an hour, something even entrenched competitors had failed to do well.

Step 2: Define the service model and resource needs

The second step is to determine the resources (for instance, vans, drivers, and service representatives) needed to meet the objective. How many employees and vans would it take for DSN to meet a promise of one-hour delivery? Prior to micromarket modeling, DSN’s managers applied—uniformly, across the network—oversimplified rules of thumb, such as a requirement that any store reach sales of $150,000 before a new van or driver would be authorized.

Micromarket modeling, building on traditional opera-tions analysis, pushed DSN to develop a more rigorous way of identifying the resources needed, at various demand levels, to fulfill a segment’s needs. To ensure that the service breakpoint could be met consistently in a variety of operating environments, the company paid particular attention to factors, such as traffic patterns and the number of customers, that limited capacity as a result of external constraints or that strained capacity during periods of peak demand. Now, instead of blindly entering the new business, the company could ask more informed questions about the trade-offs involved in allocating specific resources.

Step 3: Map the financial implications at the local level

The final step in the micromarket model is to map the financial implications of various resource configurations. For this purpose, DSN needed to have a clear understanding of the cost structure of the new delivery business and the likely combination of resources required at each location. By charting the cost of the different resource options as a function of a range of potential sales levels, the company generated a set of analytical ("iso-resource") curves¹ that show the operating margins associated with a fixed set of resources over a range of volumes (Exhibit 2, part 1).

Overlaying the resource requirements for meeting a given level of service—say, 80 percent on-time performance—produces the service-performance curve associated with it (Exhibit 2, part 2). The curve has a jagged shape because adding resources to maintain a consistent level of service as demand goes up makes the financial performance of the business "jump" from one iso-resource curve to another.

Exhibit 3 shows that a store can guarantee 80 percent on-time performance in stores with sales of up to $200,000 and of up to $450,000 with one and two vans, respectively; at the higher sales volume, the store must either add another van or accept on-time performance of less than 80 percent. Drawing a distinct curve for each service strategy generates a series of charts that identify the financial implications of different choices.

With the help of these curves, DSN realized that meeting a one-hour delivery-performance level could produce sufficient profits to justify the new business, provided that each location reached a sufficient volume level. At certain volume levels, each location could begin to exploit scale economies. Since each store’s service representative was at first underutilized, for example, increasing sales to the point at which that employee was completely busy cut the company’s customer service costs per unit, and as more vans were added within a fixed geography, the stores could chart more efficient delivery routes that reduced the time needed for deliveries. DSN also identified the profit "dead zone," at which additional resources actually had a large negative impact on store-level returns because the associated sales growth required to compensate for the additional resource was unattainable.

The implication of the micromarket-modeling approach was thus that DSN should pursue the new business only at stores that had enough high-value commercial customers within their immediate vicinity to reach the scale threshold. Micromarket modeling was used to determine that the optimal service area for each store had a five-mile radius, an area that best balanced volume and margin trade-offs. The company redrew the boundaries of its service territories and provided preferential service to customers only within the five-mile limit. Beyond it, the stores kept only the most valuable customers, which received limited service. Stores that needed unreasonably high market-share levels to meet the hurdle would be excluded from the new business. DSN has set ambitious targets—quadrupling revenues and tripling earnings margins within three years—for stores pursuing the commercial market. In some of the pilot regions, the company expects specific stores to move from bottom- to top-quartile performance in just a few months.

Thanks to this new approach, DSN could develop customized local strategies in two additional pilot markets. One, in an urban area, had very high customer den-sities, which forced the company to open a larger number of retail locations than its corporate guidelines allowed. In a more rural market, the intensity of competition and the low density of customers offered a much less appealing proposition, but thanks to the micromarket-modeling approach, DSN set less aggressive targets and reduced the likelihood of poor performance by investing resources at lower, more appropriate levels. The analysis also helped DSN to evaluate rural markets and to invest only in those with higher potential.

A new way to manage distributed service networks

In instances where resources must be invested in advance of sales, micromarket modeling can be employed to replace gut instinct

The example of DSN contrasts sharply with the way many companies pursue new business opportunities for distributed networks (Exhibit 4). Making investment decisions in the top-down manner too often leads to mistakes. A store with a low-potential local market might not be in a position to succeed, for example, but a manager might nonetheless continue to add resources there in an attempt to improve its sales volumes, in the process denying another store resources that could have been used more profitably. By contrast, DSN could tell—from market to market and location to location—whether the new model would be profitable, and the company based its decisions on that very granular level of analysis. Where resources must be invested ahead of sales, as in this case, micromarket modeling can be a powerful tool to replace gut instinct with a more informed understanding of the sales levels that can generate attractive financial returns.

Building on insights gained at individual locations, managers can use micromarket modeling to make two broad kinds of decisions: determining the appropriate number, geographic coverage, and staffing levels of new and existing outlets within a market and establishing a framework for comparing markets to set priorities for new investments and rewarding performance. Consider a few such decisions.

For the regional manager of an existing service network, defining the appropriate service territory of each location is critical. Using assumptions about sales and constraints on service levels (such as 80 percent on-time performance), such a manager can identify the resources required to expand or reduce a territory, draw a set of service-performance curves for each option, and determine the highest realistic level of profit.

The same methodology can be used to choose new locations in an existing market: the expected benefit from them can be quantified and compared with the impact of lower volume levels at existing locations, which may approach the dead zones of the business model. Suppose that a national coffee chain is debating whether to add another store in a city or to branch out to a new one. Micromarket modeling can help provide the answer by clarifying the financial implications of a new store in the existing market, the resulting cannibalization of the old stores, and the optimal coverage of all these locations.

Meanwhile, the corporate center can use micromarket modeling to make networkwide decisions about entry into new markets, the allocation of capital, growth targets, and performance rewards—thus making the company better able to customize its local strategies. Market-level curves² can be used to compare regions and to make decisions about new investments and targets for growth. Companies can evaluate the allocation of corporate overhead—the addition of an assistant regional manager, for example, or of a secondary warehouse—much as they evaluate the allocation of resources on the store level. A clothing retailer could use this approach to determine which markets it should invest in and where it should locate each of its stores.

When financial targets for markets are linked with information about customers and competitors, managers can move away from generic corporate targets and toward tailored local strategies—a potential performance breakthrough for distributed service networks. The manager of a retail bank chain, for example, could develop differentiated growth targets for the various branches, setting more ambitious goals for the stronger ones and helping the whole region achieve an aggregate target of 10 percent. Most local markets require different overall combinations of capital investment, improvements in service levels, and competitive responses. But grouping individual types of locations into peer groups as defined by the characteristics of particular local customers or competitors (such as urban or rural, chains or independents) can help regional managers make better-informed decisions about subjects ranging from the performance of individual locations to the management of incentives.

Micromarket modeling fills an analytical void that plagues managers of growing service networks. Using relatively simple and available tools, this approach can provide previously inaccessible insights into the development of winning strategies for service businesses. It can help senior managers to make strategic decisions, such as whether new businesses should be created, and to devise tailored, market-specific strategies. And it can help managers work closer to their customers and improve their operational and strategic decisions both at individual locations and at the regional level.