Purpose: Key indicator of market competitiveness.

Market share is an indicator of how well a firm is doing against its competitors. This metric, supplemented by changes in sales revenue, helps managers evaluate both primary and selective demand in their market. That is, it enables them to judge not only total market growth or decline but also trends in customers’ selections among competitors. Generally, sales growth resulting from primary demand (total market growth) is less costly and more profitable than that achieved by capturing share from competitors. Conversely, losses in market share can signal serious long-term problems that require strategic adjustments. Firms with market shares below a certain level may not be viable. Similarly, within a firm’s product line, market share trends for individual products are considered early indicators of future opportunities or problems.

Construction

Market Share: The percentage of a market accounted for by a specific entity.

Unit Market Share: The units sold by a particular company, as a percentage of total market sales, measured in the same units.

Unit Market Share (%)=Unit Sales (#)Total Market Unit Sales (#)

This formula, of course, can be rearranged to derive either unit sales or total market unit sales from the other two variables, as illustrated in the following:

Unit Sales(#)=Unit Market Share (%)*Total Market Unit Sales(#)Total Market Unit Sales (#)=Unit Sales(#)Unit Market Share(%)

Revenue Market Share: This differs from Unit Market Share in that it reflects the prices at which goods are sold. In fact, a relatively simple way to calculate relative price is to divide revenue market share by unit market share (see Section 8.1).

Revenue Market Share(%)=Sales Revenue($)Total Market Sales Revenue($)

As with the unit market share, this equation for revenue market share can be rearranged to calculate either sales revenue or total market sales revenue from the other two variables.

Data Sources, Complications, and Cautions

Market definition is never a trivial exercise: If a firm defines its market too broadly, it may dilute its focus. If it does so too narrowly, it will miss opportunities and allow threats to emerge unseen. To avoid these pitfalls, as a first step in calculating market share, managers should define the served market in terms of unit sales or revenues for a specific list of competitors, products, sales channels, geographic areas, customers, and time periods (for example, “Among grocery stores, we are the revenue market share leader in sales of frozen Italian food entrées in the Northeastern U.S.”).

Data parameters must be carefully defined: Although market share is likely the single most used marketing metric, there is no generally acknowledged best method for calculating it. This is unfortunate, as different methods may yield not only different computations of market share at a given moment but also widely divergent trends over time. The reasons for these disparities include variations in the lenses through which share is viewed (units versus dollars), where in the channel the measurements are taken (shipments from manufacturers versus consumer purchases), market definition (scope of the competitive universe), and measurement error. In the situation analysis that underlies strategic decisions, managers must be able to understand and explain these variations.

We illustrate the complexities involved in quantifying market share by looking at the competitive dynamics in the automobile industry and at General Motors in particular:

“With market share sliding in the first two months of the year, from 27.2% to 24.9%—the lowest level since a two-month strike shut the company down in 1998—GM as a whole expects a net loss of $846 million the first quarter.”²

Reviewing this statement, drawn from Business Week in 2005, a marketing manager might immediately pose a number of questions:

• Do these figures represent unit (autos sold) or revenue (dollar) market shares?

• Does this trend hold for both unit and revenue market shares at GM?

• Was revenue market share calculated before or after rebates and discounts?

• Do the underlying sales data reflect factory shipments, which relate directly to the manufacturer’s current income statement, or sales to consumers, which are buffered by dealer inventories?

• Does the decline in market share translate to an equivalent percentage decrease in sales, or has the total market size changed?

Managers must determine whether a stated market share is based on shipment data, channel shipments, retail sales, customer surveys, or some other source. On occasion, share figures may represent combinations of data (a firm’s actual shipments, for example, set against survey estimates of competitors’ sales). If necessary, managers must also adjust for differences in channels.

The time period measured will affect the signal-to-noise ratio: In analyzing short-term market dynamics, such as the effects of a promotion or a recent price change, managers may find it useful to measure market share over a brief period. Short-term data, however, generally carry a low signal-to-noise ratio. By contrast, data covering a longer period will be more stable but may obscure important recent changes in the market. Applied more broadly, this principle also holds in aggregating geographic areas, channel types, or customers. When choosing markets and time periods for analysis, managers must optimize for the type of signal that is most important.

Potential bias in reported shares: One way to find data for market sizing is through surveys of customer usage (see Section 2.7). In interpreting these data, however, managers must bear in mind that shares based on reported (versus recorded) sales tend to be biased toward well-known brands.

Related Metrics and Concepts

Served Market: The portion of the total market for which a firm competes. This may exclude geographic regions or product types. In the airline industry, for example, as of early 2020, Ryanair did not fly to the United States. Consequently, the United States would not be considered part of its served market.

Purpose: To assess a firm’s or a brand’s success and its position in the market.

A firm with a market share of 25% would be a powerful leader in many markets but a distant number two in others. Relative market share offers a way to benchmark a firm’s or a brand’s share against that of its largest competitor, enabling managers to compare relative market positions across different product markets. Relative market share gains some of its significance from studies—albeit controversial ones—suggesting that major players in a market tend to be more profitable than their competitors. This metric was further popularized by the Boston Consulting Group (BCG) in its famous matrix of relative share and market growth (see Figure 2.1).

In the BCG matrix, one axis represents relative market share—a surrogate for competitive strength. The other represents market growth—a surrogate for potential. Along each dimension, products are classified as high or low, and each is placed in one of four quadrants. In the traditional interpretation of this matrix, products with high relative market shares in growing markets are deemed stars, suggesting that they should be supported with vigorous investment. The cash for that investment may be generated by cash cows, products with high relative shares in low-growth markets. Problem child products may have potential for future growth but hold weak competitive positions. Finally, dogs have neither strong competitive position nor growth potential.

Construction

Relative Market Share(I)=Brand's Market Share (%)Brand's Largest Competitor's Market Share(%)

Relative market share can also be calculated by dividing brand sales (#, $) by largest competitor’s sales (#, $) because the common factor of total market sales (or revenue) cancels out.

Related Metrics and Concepts

Market Concentration: The degree to which a relatively small number of firms accounts for a large proportion of the market. This is also known as the concentration ratio. It is often calculated for the largest three or four firms in a market. The eight-firm concentration ratio is also popular.³

Three- (Four-) Firm Concentration Ratio: A metric that is the total (sum) of the market shares held by the leading three (four) competitors in a market.

Herfindahl Index: A market concentration metric derived by adding the squares of the individual market shares of all the players in a market. As a sum of squares, this index tends to rise in markets dominated by large players.

Data Sources, Complications, and Cautions

As ever, appropriate market definition and the use of comparable figures are vital prerequisites for developing meaningful results.

Related Metrics and Concepts

Market Share Rank: The ordinal position of a brand in its market, when competitors are arranged by size, with 1 being the largest.

Share of Category: This metric is derived in the same manner as market share but is used to denote a share of market within a certain retailer or class of retailers (for example, mass merchandisers).

Purpose: To understand the relative performance of a brand or category within specified customer groups.

The brand and category development indexes help identify strong and weak segments (usually demographic or geographic) for particular brands or categories of goods and services. For example, by monitoring the category development index (CDI), marketers might determine that Midwesterners buy twice as many country music CDs per capita as Americans in general, while consumers living on the East Coast buy fewer than the national average. This would be useful information for targeting the launch campaign for a new country music performer. Conversely, if managers found that a particular product had a low brand development index in a segment that carried a high CDI for its category, they might ask why that brand suffered relatively poor performance in such a promising segment.

Construction

Brand Development Index—BDI (I): An index of how well a brand performs within a given market group, relative to its performance in the market as a whole.

Brand Development Index nnnBDI(I)=Brand Sale to Group (#,$)/Householdsin Group(#)Total Catgory Sales(#,$)/Total Households(#)

The brand development index (BDI) is a measure of brand sales per person or per household within a specified demographic group or geography, compared with its average sales per person or household in the market as a whole. To illustrate its use, one might hypothesize that sales per capita of Ben & Jerry’s brand ice cream would be greater in the brand’s home state, Vermont, than in the rest of the country. By calculating Ben & Jerry’s BDI for Vermont, marketers could test this hypothesis quantitatively.

Category Development Index—CDI: An index of how well a category performs within a given market segment, relative to its performance in the market as a whole.

Category Development Index(I)=Category Sales to Group(#,$)⁢/Householdsin Group(#)Total Category Sales(#,$)⁢/Total Households(#)

Similar in concept to the BDI, the CDI demonstrates where a category shows strength or weakness relative to its overall performance. By way of example, Boston enjoys high per-capita consumption of ice cream. Bavaria and Ireland both show higher per-capita consumption of beer than Iran.

Data Sources and Complications

In calculating BDI or CDI, a precise definition of the segment under study is vital. Segments are often bounded geographically, but they can be defined in any way for which data can be obtained.

Related Metrics and Concepts

CDI has been applied to retail organizations. In such an application, it measures the extent to which a retailer emphasizes one category versus others.

Category Development Index(I)=Retailer's Share of Category Sales(%)Retailer's Total Share of Market(%)

The use of CDI to assess category emphasis by particular retailers is very similar to the category performance ratio (see Section 7.1).

Construction

Penetration: The proportion of people (households) in the relevant population who bought (at least once in the period) a specific brand or a category of goods.

Market Penetration(%)=Customers Who Have Purchased a Product in the Category(#)Total Population(#)Brand Penetration(%)=Customers Who Have Purchased the Brand(#)Total Population(#)

Whereas market share focuses on the sales of a product (either units or dollars), penetration focuses on the number of buyers.

Penetration share can be thought of as the share of category households that bought the brand. Thus, a brand’s penetration share can never be greater than one. When households buy multiple brands, however, penetration shares will sum to more than one.

Penetration Share(%)=Brand Penetration(%)Market Penetration(%)

Data Sources, Complications, and Cautions

The time period over which a firm measures penetration will have a significant impact on the result. For example, even among the most popular detergent brands, many are not purchased weekly. As the time period used to define penetration becomes shorter, brand penetration declines. In contrast, although penetration share will be more volatile for shorter periods, it will not necessarily be lower.

Related Metrics and Concepts

Active customers: Customers (people, households, accounts) who have purchased the brand (category) in the current time period. When these are counted, this gives Total Number of Active Customers.

Note that any count of active customers will likely be less than the total number of customers because the latter includes customers who have bought previously but not in the current period. This is discussed in more detail in Section 5.1 (recency).

Active customers may also be expressed as a percentage of the total population. To do so divide the number of active customers by the total number of people in the population. (When assessed at a brand level, this percentage is equivalent to brand penetration).

Ever-tried: Customers who have purchased the brand at any time (see Section 4.1 for more on trial.) This is equivalent to the term “penetration.” Ever-tried customers who are not active are sometimes referred to as former customers (if there is little chance they will buy in the future) or simply inactive customers (if there is a good chance they will buy in the future.) Ever-tried may be expressed as a percentage of the population. To do so divide the number of customers who have ever tried by the total number of customers in the population.

Acceptors/Accepters: Customers who research indicates are willing to buy the brand; the opposite of rejectors.

Purpose: To understand the source of market share in terms of breadth and depth of consumer franchise as well as the extent of relative category usage (heavy users/larger customers versus light users/smaller customers).

Construction

Share of Requirements: A given brand’s share of purchases in its category, measured solely among customers who have purchased that brand. Also known as share of wallet.

When calculating share of requirements, marketers may use either dollars or units.

Unit Share of Requirments(%)=Brand Purchases(#)Total Category Purchases by Brand Buyers(#)Revenue Share of Requirments(%)=Brand Purchases($)Total Category Purchases by Brand Buyers($)

The best way to think about share of requirements is as the market share enjoyed by a brand among the customers who buy it.

Data Sources, Complications, and Cautions

Double Jeopardy: As mentioned earlier, the three components of market share do not move independently in practice. In an empirical observation labeled “double jeopardy,” brands with lower market share almost invariably have lower penetration share and lower share of requirements. The name “double jeopardy” captures the notion that low-share brands are punished twice for their lower share. Not only do fewer households buy them but also those buying households buy less of them (lower share of requirements).

Double jeopardy is a very real and pervasive phenomenon. One implication is that improvements in market share will be accompanied by improvements in both penetration and share of requirements. It is also true that brands appear to vary more with respect to penetration than share of requirements—which has led others to argue that it is better to try to increase penetration (sell to more households) than to increase the loyalty of the brand’s current customers. We do not endorse that argument.

One explanation for double jeopardy is that low-share brands do not get broad distribution. Thus, it is more difficult for the few customers who prefer them to find and buy them, and share of requirements suffers. In this way of thinking, it is market share that leads to both penetration and share of requirements.

There is also a purely statistical explanation for double jeopardy that we will describe at the end of Section 2.6.

Related Metrics and Concepts

Sole Usage Percentage: The proportion of a brand’s customers who bought only that brand’s product and did not buy from competitors. Sole users are a combination of extremely loyal customers, customers with limited access to brands, and customers who just happened to buy one unit during the period. Sole users are also known as purely loyals. Among sole users, a brand’s share of requirement is, by definition, 100%.

Sole Usage(%)=Customers Who Buy Only the Brand in Question(#)Total Brand Customers(#)

Number of Brands Purchased: During a given period, some customers may buy only a single brand within a category, whereas others buy two or more. In evaluating loyalty to a given brand, marketers sometimes compare the average number of brands purchased by brand customers to the average number purchased by all customers in the category.

Repeat Rate: The percentage of brand customers in a given period who are also brand customers in the subsequent period.

Repurchase Rate: The percentage of customers for a brand who repurchase that brand on their next purchase occasion.

Confusion abounds in this area. In these definitions, we have tried to distinguish a metric based on calendar time (repeat rate) from one based on “customer time” (repurchase rate). In Chapter 5, “Customer Profitability,” we will describe a related metric, retention, which is used in contractual situations in which the first non-renewal (non-purchase) signals the end of a customer relationship. Although we suggest that the term retention be applied only in contractual situations, you will often see repeat rates and repurchase rates referred to as retention rates. Due to a lack of consensus on the use of these terms, marketers should not rely on the names of these metrics as perfect indicators of how they are calculated.

As with penetration, the interpretation of repeat rate depends on the time period covered. The shorter the time period, the lower will be the repeat rate. One minus the repeat rate is sometimes called the turnover rate.

Purpose: To define and measure whether a firm’s consumers are “heavy users.”

The usage index answers the question “How heavily do our customers use the category of our product?”

Construction

Usage Index: The ratio of the average category usage for the customers of a brand to the average category usage of all customers.

The usage index can be calculated on the basis of units or dollars.

Usage Index(I)=Average Total Purchase in Category byBrand Customers(#,$)Average Total Purchases in Category All Customers forThatCategory(#,$)

As previously noted, market share can be calculated as the product of three components: penetration share, share of requirements, and usage index (see Section 2.4). Consequently, we can calculate a brand’s usage index if we know its market share, penetration share, and share of requirements, as follows:

Usage Index(I)=Market Share(%)Penetration Share (%)*Share of Requirements(%)

This equation works for market shares defined in either units or dollars, as long as the construction of share of requirements and usage index matches. Comparing a brand’s dollar usage index to its unit usage index, marketers can determine whether the brand purchasers pay more (less) per category unit.

Data Sources, Complications, and Cautions

The usage index does not indicate how heavily customers use a specific brand, only how heavily they use the category. A brand can have a high usage index, for example, meaning that its customers are heavy category users, even if those customers use the brand in question to meet only a small share of their needs.

Understanding Double Jeopardy and Usage Index

In the previous section we described an empirical generalization called double jeopardy. Brands with lower share are punished twice in that they sell to fewer people (lower penetration) who buy them less frequently (lower share of requirements). Before jumping to conclusions about the marketing strategy implications of double jeopardy, however, it is important to understand the statistical explanation for double jeopardy.

Double jeopardy is a natural consequence of two things:

• How the component metrics are defined—with customers of a brand being defined as all households that bought the brand during the period.

• How households behave—as households often buy multiple brands within a given period for inexplicable reasons.

To illustrate, consider a market with three brands (A, B, and C) serving ten households with unit sales for the period given in Table 2.4.

The category summary metrics paint a clear picture of this ten-household market. All households were loyal to a single brand, and all households bought the same number of units. The 60/30/10 share split exactly matches the brands’ penetrations (PENs) of 60%, 30%, and 10%, respectively (see Table 2.5).

But now let us add an 11th household that is a heavy, but indiscriminate, buyer purchasing 12 units of A, 6 units of B, and 2 units of C (see Table 2.6).

We contend that the presence of this multiple-brand buyer doesn’t really change our view of the relative loyalties enjoyed by the three brands. Brand C got a 10% share of the ten loyal households and also a 10% share of not-so-loyal household 11. Brand C performed exactly as one might expect.

But look at the new category summary statistics shown in Table 2.7.

The penetration metrics have all gone up a bit and now sum to something greater than one because household 11 is counted as a customer of each of the three brands. A consequence of this double counting and the fact that household 11 was a heavy category user is that all three usage indices are greater than one. In our experience, this is often the case. Heavy-using households tend to buy several brands (which makes sense), and we end up with usage indices that are all greater than one.

It also makes sense that the presence of household 11 causes share of requirements to decrease for all three brands. But the important observation is that the decrease in share of requirements is much greater for the small-share brand simply because it is the small-share brand, and heavy-user household 11 has a bigger influence on its summary statistics. (Household 11 is 1/2 of brand C’s customers but only 1/7 of brand A’s.) The result is double jeopardy caused by the presence of household 11 and the fact that household 11 gets counted as a customer of all three brands. Although this is appropriate in one sense (household 11 is a customer for all three brands) and allows our market share decomposition to hold, it affects the behavior of share of requirements and usage index metrics in a way that makes them challenging to interpret. Certainly for us, share of requirements is a less-than-perfect measure of the concept of loyalty.

Related Metrics and Concepts

See also the discussion of brand development index (BDI) and category development index (CDI) in Section 2.3.

Purpose: To track trends in customer attitudes and behaviors.

Awareness, attitudes, and usage (AAU) metrics relate closely to what has been called the Hierarchy of Effects, an assumption that customers progress through sequential stages from lack of awareness, through initial purchase of a product, to brand loyalty (see Figure 2.2). AAU metrics are generally designed to track these stages of knowledge, beliefs, and behaviors. AAU studies also may track “who” uses a brand or product—in which case customers are defined by category usage (heavy/light), geography, demographics, psychographics, media usage, and whether they purchase other products.

Information about attitudes and beliefs offers insight into the question of why specific users do or do not favor certain brands. Typically, marketers conduct surveys of large samples of households or business customers to gather these data.

Construction

Awareness, attitudes, and usage studies feature a range of questions that aim to shed light on customers’ relationships with a product or brand (see Table 2.8). For example, who are the acceptors and rejecters of the product? How do customers respond to a replay of advertising content?

Marketers use answers to these questions to construct a number of metrics. Among these, certain “summary metrics” are considered important indicators of performance. In many studies, for example, customers’ “willingness to recommend” and “intention to purchase” a brand are assigned high priority. Underlying these data, various diagnostic metrics help marketers understand why consumers may be willing—or unwilling—to recommend or purchase that brand. Consumers may not have been aware of the brand, for example. Alternatively, they may have been aware of it but did not subscribe to one of its key benefit claims.

Data Sources, Complications, and Cautions

Sources of AAU data include

• Warranty cards and registrations, often using prizes and random drawings to encourage participation.

• Regularly administered surveys, conducted by organizations that interview consumers via telephone, mail, web, or other technologies, such as handheld scanners.

Even with the best methodologies, however, variations observed in tracking data from one period to the next are not always reliable. Managers must rely on their experience to distinguish seasonality effects and “noise” (random movement) from “signal” (actual trends and patterns). Certain techniques in data collection and review can also help managers make this distinction:

• Adjust for periodic changes in how questions are framed or administered. Surveys can be conducted via mail or telephone, for example, among paid or unpaid respondents. Different data-gathering techniques may require adjustment in the norms used to evaluate a “good” or “bad” response. If sudden changes appear in the data from one period to the next, marketers are advised to determine whether methodological shifts might play a role in this result.

• Try to separate customer from non-customer responses; they may be very different. Causal links among awareness, attitudes, and usage are rarely clear-cut. Though the hierarchy of effects is often viewed as a one-way street, on which awareness leads to attitudes, which in turn determine usage, the true causal flow might also be reversed. When people own a brand, for example, they may be predisposed to like it.

• Triangulate customer survey data with sales revenue, shipments, or other data related to business performance. Consumer attitudes, distributor and retail sales, and company shipments may move in different directions. Analyzing these patterns can be challenging but can reveal much about category dynamics. For example, toy shipments to retailers often occur well in advance of the advertising that drives consumer awareness and purchase intentions. These, in turn, must be established before retail sales. Adding further complexity, in the toy industry, the purchaser of a product might not be its ultimate consumer. In evaluating AAU data, marketers must understand not only the drivers of demand but also the logistics of purchase.

• Separate leading from lagging indicators whenever possible. In the auto industry, for example, individuals who have just purchased a new car show a heightened sensitivity to advertisements for its make and model. Conventional wisdom suggests that they’re looking for confirmation that they made a good choice in a risky decision. By helping consumers justify their purchase at this time, auto manufacturers can strengthen long-term satisfaction and willingness to recommend.

Related Metrics and Concepts

Likeability: Because AAU considerations are so important to marketers, and because there is no single “right” way to approach them, specialized and proprietary systems have been developed. Of these, one of the best known is the Q Scores rating of “likeability.” A Q Score is derived from a general survey of selected households, in which a large panel of consumers share their feelings about brands, celebrities, and television shows.⁴

Q Scores rely upon responses reported by consumers. Consequently, although the system used is sophisticated, it is dependent on consumers understanding and being willing to reveal their preferences.

Segmentation by Geography, or Geo-clustering: Marketers can achieve insight into consumer attitudes by separating their data into smaller, more homogeneous groups of customers. One well-known example of this is Prizm. Prizm assigns U.S. households to clusters based on zip code,⁵ with the goal of creating small groups of similar households. The typical characteristics of each Prizm cluster are known, and these are used to assign a name to each group. “Golden Ponds” consumers, for example, comprise elderly singles and couples leading modest lifestyles in small towns. Rather than monitor AAU statistics for the population as a whole, firms often find it useful to track these data by cluster.

Purpose: Customer satisfaction provides a leading indicator of consumer purchase intentions and loyalty.

Customer satisfaction data are among the most frequently collected indicators of market perceptions. Their principal use is twofold:

• Within organizations, the collection, analysis, and dissemination of these data send a message about the importance of tending to customers and ensuring that they have positive experiences with the company’s goods and services.

• Although sales or market share can indicate how well a firm is performing currently, satisfaction is perhaps the best indicator of how likely it is that the firm’s customers will make further purchases in the future. Much research has focused on the relationship between customer satisfaction and retention. Studies indicate that the ramifications of satisfaction are most strongly realized at the extremes. On the scale in Figure 2.3, individuals who rate their satisfaction level as “5” are likely to become return customers and might even evangelize for the firm. Individuals who rate their satisfaction level as “1,” by contrast, are unlikely to return. Further, they can hurt the firm by making negative comments about it to prospective customers. Willingness to recommend is a key metric related to customer satisfaction.

Construction

Customer Satisfaction: The number of customers, or percentage of total customers, whose reported experience with a firm, its products, or its services (ratings) exceeds specified satisfaction goals.

Willingness to Recommend: The percentage of surveyed customers who indicate that they would recommend a brand to friends.

These metrics quantify an important dynamic. When a brand has loyal customers, it gains positive word-of-mouth marketing, which is both free and highly effective.

Customer satisfaction is measured at the individual level, but it is almost always reported at an aggregate level. It can be, and often is, measured along various dimensions. A hotel, for example, might ask customers to rate their experience with its front desk and check-in service, with the room, with the amenities in the room, with the restaurants, and so on. In addition, in a holistic sense, the hotel might ask about overall satisfaction “with your stay.”

Customer satisfaction is generally measured on a five-point scale (see Figure 2.4).

Satisfaction levels are usually reported as either “top box” or, more likely, “top two boxes.” Marketers convert these expressions into single numbers that show the percentage of respondents who checked either a “4” or a “5.” (This term is the same as that commonly used in projections of trial volumes; see Section 4.1.)

Regardless of the scale used, the objective is to measure customers’ perceived satisfaction with their experience of a firm’s offerings. Marketers then aggregate these data into a percentage of top-box responses.

In researching satisfaction, firms generally ask customers whether their product or service has met or exceeded expectations. Thus, expectations are a key factor behind satisfaction. When customers have high expectations and the reality falls short, they will be disappointed and will likely rate their experience as less than satisfying. For this reason, a luxury resort, for example, might receive a lower satisfaction rating than a budget motel—even though its facilities and service would be deemed superior in “absolute” terms.

Data Sources, Complications, and Cautions

Surveys constitute the most frequently used means of collecting satisfaction data. As a result, a key risk of distortion in measures of satisfaction can be summarized in a single question: Who responds to surveys?

Response bias is endemic in satisfaction data. Disappointed or angry customers often welcome a means to vent their opinions. Contented customers often do not. Consequently, although many customers might be happy with a product and feel no need to complete a survey, the few who had a bad experience might be disproportionately represented among respondents. Most hotels, for example, place response cards in their rooms, asking guests, “How was your stay?” Only a small percentage of guests ever bother to complete those cards. Not surprisingly, those who do respond probably had a bad experience. For this reason, marketers can find it difficult to judge the true level of customer satisfaction. By reviewing survey data over time, however, they may discover important trends or changes. If complaints suddenly rise, for example, that may constitute early warning of a decline in quality or service. (See number of complaints in the following section.)

Sample selection may distort satisfaction ratings in other ways as well. Because only customers are surveyed for customer satisfaction, a firm’s ratings may rise artificially as deeply dissatisfied customers take their business elsewhere. Also, some populations may be more frank than others or more prone to complain. These normative differences can affect perceived satisfaction levels. In analyzing satisfaction data, a firm might interpret rating differences as a sign that one market is receiving better service than another, when the true difference lies only in the standards that customers apply. To correct for this issue, marketers are advised to review satisfaction measures over time within the same market.

A final caution: Because many firms define customer satisfaction as “meeting or exceeding expectations,” this metric may fall simply because expectations have risen. Thus, in interpreting ratings data, managers may come to believe that the quality of their offering has declined when that is not the case. Of course, the reverse is also true. A firm might boost satisfaction by lowering expectations. In so doing, however, it might suffer a decline in sales as its product or service comes to appear unattractive.

Related Metrics and Concepts

Trade Satisfaction: Founded on the same principles as consumer satisfaction, measures the attitudes of trade customers.

Number of Complaints: The number of complaints lodged by customers in a given time period.

Purpose: To measure how well the brand or company is succeeding in creating satisfied, loyal customers.

Net Promoter Score®⁷ (NPS) is a registered trademark of Frederick R. Reichheld, Bain & Company, and Satmetrix that is a particularly simple measure of the satisfaction/loyalty of current customers. Customers are surveyed and asked (on a ten-point scale) how likely they are to recommend the company or brand to a friend or colleague. Based on their answers to this single question, customers are divided into several categories:

• Promoters: Customers who are willing to recommend the company to others (who gave the company a rating of “9” or “10”).

• Passives: Satisfied but unenthusiastic customers (ratings of “7” or “8”).

• Detractors: Customers who are unwilling to recommend the company to others (ratings of “0” to “6”).

High NPSs generally mean that a company is doing a good job of securing its customers’ loyalty and active evangelism. Low and negative NPSs are important early warning signals for the firm. Because the metric is simple and easy to understand, it provides a stable measure companies use to motivate employees and monitor progress.

Construction

The Net Promoter Score (NPS) is created by subtracting the percentage of detractors among current customers from the percentage of promoters among current customers.

Net Promoter Score (I) = Percentage of Promoters (%) - Percentage of Detractors (%)

For example, if a survey of a company’s customers reports that there were 20% promoters, 70% passives, and 10% detractors, the company would have an NPS of 20 − 10 =10.

Data Sources, Complications, and Cautions

Although the trademarked NPS asks a specific question, uses a scale up to 10, and defines promoters, passives, and detractors in a particular way, it is easy to imagine other versions of NPS-like metrics that differ with respect to the wording of the question, the scale used (1–5 rather than 0–10), and the definitions (and labels) of the resulting groups of responders. The defining features of NPS are that it is constructed from responses to a question about willingness to recommend and is a net measure found by subtracting the fraction unwilling to recommend from the fraction willing to recommend and leaving out those in the middle.

The same NPS can indicate different business circumstances. For instance, an NPS of 0 can indicate highly polarized customers with 50% promoters and 50% detractors, or it can indicate a totally ambivalent customer base with 100% passives. Getting the NPS may be a good way of starting a discussion about customer perceptions of a brand. As it is an average of current customers’ responses, managers must drill down to the data to understand the precise situation their business faces.

This score in specific circumstances can generate results that could mislead a manager who is not being careful. For example, consider a company whose current customers are 30% promoters, 30% detractors, and 40% passives. This company’s NPS is an unimpressive 0 because 30% − 30% = 0.

Now suppose that a new competitor steals two-thirds of the company’s detractors, and because these detractors immediately defect to the new competitor, they cease to be customers of the company. The NPS is remeasured.

Promoters are now 30% / (100% − 20% = 80%) = 37.5% of the customers that remain.

Passives are now 40% / (100% − 20% = 80%) = 50% of the customers that remain.

Detractors are now only (30% − 20% = 10%) / (100% − 20% = 80%) = 12.5% of the customers that remain.

The NPS is now 37.5% − 12.5% = a very healthy-looking 25.

The defection of the most vulnerable and unhappy customers led directly to an increase in NPS. Managers should make sure they fully understand what has happened.

While benchmarking is often a useful exercise, it is inappropriate to directly apply this measure across categories. Some products are in categories that are more likely to gain engagement, both positive and negative, than others.

A high Net Promoter Score, while generally desirable, does invite the question whether the company is properly monetizing the value it is providing to the consumer. The easiest way to develop a high NPS is to provide a highly valued product to customers for free. Why wouldn’t they be happy to recommend you? While there might be strategic reasons for situations like this to be acceptable to the company in the short or medium term, it probably wouldn’t be a viable long-term strategy.

The NPS is calculated from survey data. As such, it may suffer from the problems common to most surveys, and the results should be interpreted in light of other data, such as sales trends. Is increased customer satisfaction leading to increased sales? If so, fine; if not, why not?

Although the NPS has received much attention and relatively rapid adoption, it has also been the target of significant criticism. Consultant Timothy Keiningham and his coauthors claim that the benefits of the measure have been overstated relative to other measures of loyalty and satisfaction.^8,9

Purpose: To assess the commitment of a firm’s or a brand’s customer base.

Brand or company loyalty is a key marketing asset. Marketers evaluate aspects of it through a number of metrics, including repurchase rate, share of requirements, willingness to pay a price premium, and other AAU measures. Perhaps the most fundamental test of loyalty, however, can be captured with a simple question: When faced with a situation in which a brand is not available, will customers search further for it (that is, go to another store), or will they substitute the best available option?

When a brand enjoys loyalty at this level, its provider can generate powerful leverage in trade negotiations. Often, such loyalty will also give providers time to respond to a competitive threat. Customers will stay with them while they address the threat.

Loyalty is grounded in a number of factors, including

• Satisfied and influential customers who are willing to recommend the brand

• Hidden values or emotional benefits, which are effectively communicated

• A strong image for the product, the user, or the usage experience

Purchase-based loyalty metrics are also affected by whether a product is broadly and conveniently available for purchase and whether customers enjoy other options in its category.

Construction

Willingness to Search: The likelihood that customers will settle for a second-choice product if their first choice is not available. Also called “accept no substitutes.”

Willingness to search represents the percentage of customers who are willing to leave a store without a product if their favorite brand is unavailable. Those willing to substitute constitute the balance of the population.

Data Sources, Complications, and Cautions

Loyalty has multiple dimensions. Consumers who are loyal to a brand in the sense of rarely switching may or may not be willing to pay a price premium for that brand or recommend it to their friends. Behavioral loyalty may also be difficult to distinguish from inertia or habit. When asked about loyalty, consumers often don’t know what they will do in new circumstances. They may not have accurate recall about past behavior, especially in regard to items with which they feel relatively low involvement.

Furthermore, different products generate different levels of loyalty. Few customers will be as loyal to a brand of matches, for example, as to a brand of baby formula. Consequently, marketers should exercise caution in comparing loyalty rates across products. Rather, they should look for category-specific norms.

Degrees of loyalty also differ between demographic groups. Older consumers have been shown to demonstrate the highest loyalty rates.

Even with these complexities, however, customer loyalty remains one of the most important metrics to monitor. Marketers should understand the worth of their brands in the eyes of the customer—and of the retailer.

Purpose: Obtain deeper insights into consumer behavior.

Consumers do things for reasons they do not fully understand and cannot articulate. Surveys only reveal what people can or care to tell us and are subject to post hoc rationalization (that is, making up a reason for actions that sounds plausible but only after the decision was made). Surveys thus have limitations. The good news is that technological advances have allowed marketers to gain greater insight into how consumers react, especially preconciously or non-consciously. We have a better understanding of how consumers view visual advertisements and even which parts of the consumer’s brain are active when deciding between alternatives. Many companies provide specialist consultancy services to help marketers benefit from these new techniques.

Current research in cognitive neuroscience suggests that decision making is less deliberate than once thought and instead relies heavily on early emotional responses. Traditional market research methods, such as focus groups and surveys, are unable to assess consumers’ initial, preconscious reactions. Consumer neuroscience methods can reveal insights into early emotional responses to help us better understand these reactions and so give a unique insight into consumer behavior.

Construction

We consider four of the most important measures used in trying to understand consumer reactions. These are classified here by the technology rather than the measure, as the choice of technology is generally the way that marketers will encounter the metrics.

Why Use Neuro-Marketing Techniques?

Brand-equity measures often capture the progress of a brand’s marketplace performance, but neuro-marketing measures have potential to tell us more about the consumers’ experiences at a detailed level. We can also hope to better understand non-conscious consumer reactions, which consumers cannot tell us about in surveys even if they wish to do so.

Where these are likely to be especially useful is where the upfront costs are modest compared to the ongoing marketing costs. Here even a modest increase in understanding can mean large benefits from the more efficiently deployed ongoing costs. A case in point may be advertising where typically media buys (such as time on TV) are a large proportion of the costs. (The cost to create a commercial is often modest compared to the cost to air it on network TV.) Anything that even marginally improves the effectiveness of an advertisement can be highly worthwhile. (For more on advertising testing, see Section 13.3 on the Gross model.)

The study’s success was driven by the fact that the study’s upfront cost was less than the ongoing benefits gained through more efficient deployment of the media spend. The 15-second advertisements give much more bang for the buck than the 30-second advertisements.

The benefits from small improvements in effectiveness get even greater when the ongoing spending increases relative to the initial costs. Table 2.10 shows the benefits that would have been gained had Heritage been committed to five times more media spend.

Data Sources, Complications, and Cautions

One of the main limitations of neuro-marketing techniques is that the machinery can be cumbersome and hard or impossible to deploy in everyday situations. While we can ask consumers to participate in these tests, their reactions will, to a certain extent, be influenced by the setting they are in. Participating in an experiment in an fMRI scanner is not (yet) the same as shopping in your local grocery store, even if some of the measuring devices are rapidly becoming smaller and more comfortable. As technology for gathering neuroscience metrics becomes less obtrusive, we might expect some of these problems to be alleviated.

Compared to other marketing metrics, neuro-marketing requires specific expertise and training in the underlying science to interpret the data. Further, although companies are accumulating studies that assess reliability and validity of the technologies and metrics, much remains to be done.

As Varan and his colleagues note, the need to assess the variance and reliability of neuro-metrics within and between firms has become increasingly apparent.¹⁵ There remains considerable variation among suppliers and uncertainty about what differences are due to lack of reliability versus what differences are driven by each firm’s unique collection methods.

Our understanding of the brain has increased considerably in recent years, but it still remains limited. Scientists may talk of areas of the brain that are associated with a certain feeling or activity. This does not mean that there is a specific spot in the brain that, if we were to find it, could tell exactly a person’s brand preference. Neither can we hope to understand a consumer’s precise willingness to pay for any good. Neuro-marketing techniques may be useful in understanding how consumers are likely to behave and so may help at an early stage of the marketing process. Table 2.11 provides a summary of neuro-marketing measures.

Further Reading

Banaji, Mahzarin R., and Anthony G. Greenwald. (2013). Blindspot: Hidden Biases of Good People, Delacorte Press.

Hendrickson, Kirk, and Kusum L. Ailawadi. (2014). “Six Lessons for In-Store Marketing from Six Years of Mobile Eye-Tracking Research,” in Dhruv Grewal, Anne L. Roggeveen, Jens NordfÄlt (Eds.), Shopper Marketing and the Role of In-Store Marketing (pp. 57–74), Emerald Group Publishing Limited.

Knutson, Rick B., G. E. Wimmer, D. Prelec, and G. Loewenstein. (2007). “Neural Predictors of Purchases,” Neuron, 53(1), 147–156.

McClure, Samuel M., Jian Li, Damon Tomlin, Kim S. Cypert, Latané M. Montague, and P. Read Montague. (2004). “Neural Correlates of Behavioral Preference for Culturally Familiar Drinks,” Neuron, 44(2), 379–387.

Teixeira, Thales, Michel Wedel, and Rik Pieters. (2012). “Emotion-Induced Engagement in Internet Video Advertisements,” Journal of Marketing Research, 49, 144–159.

Trabulsi, Julia, Manuel Garcia-Garcia, and Michael E. Smith. (2015). “Consumer Neuroscience: A Method for Optimizing Marketing Communication,” Journal of Cultural Marketing Strategy, 1(1), 80–89.

Varan, Duane, Annie Lang, Patrick Barwise, Rene Weber, and Steven Bellman. (2015). “How Reliable Are Neuromarketers’ Measures of Advertising Effectiveness: Data from Ongoing Research Holds No Common Truth Among Vendors,” Journal of Advertising Research, 55(2), 176–191.

Venkatraman, Vinod, Angelika Dimoka, Paul A. Pavlou, Khoi Vo, William Hampton, Bryan Bollinger, Hal E. Hershfield, Masakazu Ishihara, and Russell S. Winer. (2015). “Predicting Advertising Success Beyond Traditional Measures: New Insights from Neurophysiological Methods and Market Response Modeling,” Journal of Marketing Research, 52(4), 436–452.

Willke, Joe, and Blake Burrus. (2013). “Making Advertising More of a Science Than an Art,” What’s Next, 1(2).