Despite their heightened attention to advertising spend, managers still have a hard time getting it precisely right. This should occasion no surprise: the task is extremely - perhaps even impossibly - difficult. There are problems everywhere: data are often unreliable, internal accounting and control systems irrelevant, performance criteria inappropriate, budgets unpredictable, time horizons for decisions unclear or unreasonable, and managerial incentives unfocused.

These problems do not remove the burden of getting it right. But they do redefine a bit the yardstick by which "right" gets measured. Greater accuracy and precision are, of course, always welcome. But the primary merit of any new approach to making decisions about advertising spend is to embed throughout an organization a consistent, shared discipline for systematically thinking through the complex dynamics of value creation and capture.

Such consistent discipline is especially important given the relatively fragile state of knowledge about the true economic value of advertising. There are, for example, a couple of facts about advertising that almost everyone knows. For much of the 1980s, ad spending boomed at a double-digit clip. Then, in the 1990s, expenditure began to decline in real terms. Among managers as well as students of advertising, two competing theories have developed to explain these conflicting trends:

1. Advertisers were overspending during the 1980s, or they are underspending today, or both.
2. Either (or both) rate of spending is right - it is simply that the nature of competition, the overall performance of the economy, and the role of advertising have changed from one decade to the next.

Ad spending has become so disconnected from economic reality that the numbers reflect only management’s willingness to open its wallet

Both theories are plausible and popular. Both are wrong. In fact, there is a third explanation: ad spending has become so disconnected from economic reality and from an understanding of how best to evaluate an ad’s "quality" - that is, its creativity - that these numbers do not reflect much other than management’s willingness to open its wallet.

New approaches to thinking about advertising that are anchored in effective metrics of creativity and in the economic reality of value creation are therefore needed. The one described here focuses on the economic half of this equation.

Based on existing market research techniques and information technology, which have eased tremendously the execution of such approaches, it provides managers with a rational, dependable tool for setting advertising budgets and monitoring advertising effectiveness. It also allows them to see how their products deliver value to individual market segments and learn how advertising can communicate this value to the consumer. Most of all, it helps to clarify the relationship between a dollar of media spend and a dollar of economic value.

Many companies are adopting bits and pieces of this approach, although to date no large advertisers use it systematically. They tend to direct their approach toward the market as a whole and not toward the multiple segments they want to reach. Even so, they have all found genuine benefit in the disciplined thought process it helps create and institutionalize.

A different model

In most companies, senior management gives each brand manager his or her ad budget for the year. The budget is generally based on last year’s, plus or minus a few percent. The brand manager then works with ad agencies and is "free" to spend the budget more or less as he or she pleases - within, of course, the sharp financial constraints imposed by the need to deliver against tough, year-to-year P&Ls.

This new approach works differently.

It should be familiar to senior managers because it resembles the standard capital budgeting process that operates outside advertising. Instead of being handed budgets from above, brand managers tell senior managers how much their budget should be. They are allowed to spend as much as they want on advertising - provided that the return on that spending can be shown to beat some internally agreed hurdle.

If you cannot prove that the dollars you spend persuade consumers to purchase your product, you should not advertise - period

There is no reason why senior man-agers should not be able to answer the economic portion of the simple question underlying all advertising: does my advertising work? If you cannot prove that the dollars you spend create economic value by persuading consumers to purchase your product, you should not advertise - period. On the other hand, if you can prove effectiveness, the answer may well be to put more spending behind successful campaigns for successful products in the right segments. The absolute size of the ad budget is not the point. Under this new way of thinking, issues of creativity aside, no ad budget is too large, provided that it produces positive economic returns on the capital invested.

This will not, of course, solve all the measurement problems associated with advertising. There is, as noted above, a substantial creative element that does not lend itself to the same kind of hard measurement that applies to financial analyses. Nor does it address the issue of long-term impact - although by logically modeling the way advertising works, it makes it easier to estimate the likely long-term effects than do current econometric approaches that rely on statistical chance. Nor is it necessarily the best way of proceeding in all circumstances. In most situations, however, it represents a substantial improvement over current practice, where the lack of proven techniques has meant little or no effort has gone into measurement.

This approach works by forcing the whole of advertising to be more accountable than it traditionally has been, and, consequently by increasing the level of confidence in spending decisions. Available resources are committed to better options. Most important of all, by laying bare the whole process, it exposes advertising strategy to examination and transforms it into a logical, step-by-step, bottom-up enterprise rather than a top-down activity vulnerable to the whims of all those in management who fancy themselves as advertising critics.

What happens now?

Financial planning drives the process of determining advertising spend for most existing brands

Exhibit 1 illustrates how most consumer companies typically plan and execute their advertising. In general, financial planning drives the process of determining advertising spend for existing brands. An overall marketing budget is often set during an annual planning routine which starts with revenue and profit targets and progresses to volume targets and pricing guidelines. A careful, managed process then allocates funds to the more tangible levers of marketing, such as trade and consumer promotion, with specific targets for, say, cases shipped or levels of redemption. Whatever is left over at the end is assigned to advertising.

A common outcome is that some major national brands - or even companies - have to do without national campaigns, since the money that is left cannot be stretched beyond spot radio campaigns in local markets. Indeed, a few well-known US brands of soaps, shampoos, conditioners, cookies, and crackers spent no money on network television in 1993.

Planning the advertising

The ad plan is often developed jointly by the marketing manager and the ad agency’s account management. Its main elements are the pri-mary message or theme of the campaign, positioning guidelines for execution, the definition of the target audience (usually expressed in broad demographic terms - for instance, upscale women 25 to 44) and indications of media vehicles to be used (such as network TV or spot radio).

There is limited understanding of the links between product positioning, the resulting perceptions, and their impact on the behavior of distinct segments

Many advertisers have general communications objectives, such as increasing or maintaining awareness, but set quantitative targets all the same. The image or positioning of a product tends to be driven by how management would like the brand to be perceived or by its heritage, rather than by an understanding of the links between positioning, the resulting perceptions, and their impact on the behavior of distinct segments. The need to adapt positioning to the differing needs of individual segments is often acknowledged, but rarely acted upon. A unified positioning can work for a brand that appeals to a specific segment - as, for example, MCI’s well-established price positioning in the US long-distance telephone market. Conversely, a single positioning probably would not work for AT&T, with its much larger base of customers. To take another example, many US banks adopt a communications strategy that involves spending the bulk of their advertising effort on attracting new customers through the mass media. Yet the communications objectives of boosting retention, building loyalty, and broadening the range of products bought by existing customers can often be more profitable.

Copy development and approval

The key decisions in the development of an ad campaign usually involve too many players

There is a tendency for key decisions in the development of an ad campaign to involve too many players, some of them ill-informed and inexperienced. Copy may be tested using syndicated services that measure a commercial’s ability to attract attention and break through the clutter. This means using yardsticks such as recall, recognition, and persuasion, but ad managers seldom calculate how well the copy actually succeeds in communicating the message and achieving the desired change in perception.

Recall and recognition scores, for example, are inappropriate when the aim of the copy is to change the audience’s perceptions because a commercial that is widely remembered and recognized by consumers may not actually succeed in changing their perceptions of the product or service being advertised. Perceptual maps or image engineering techniques, which would be appropriate for measuring changes in perception, are used occasionally to help determine positioning strategy, but rarely to track changes in perception or to test advertising copy.

One automobile manufacturer that used focus groups to test its copy was unable to distinguish between the impact of different versions of copy on target customer segments. Despite a high level of spending, the company had no real impact on the levers that mattered to its customers, and so it exercised only a minimal influence over their buying intentions.

Execution

The weakest link in the advertising process is in the delivery of the message to the target audience

The weakest link in the advertising process is in the delivery of the message to the target audience. Although it is possible to pinpoint an audience or readership with great accuracy, advertisers, agencies, and media have traditionally been content to buy and sell on the basis of broad demographic groupings.

Agencies and advertisers will often expand their original precise definitions of their target audience to correspond to the more broadly defined television or magazine audiences. Despite the waste, this can sometimes be justified, since it is hard to make the numbers add up with a splintered media strategy. Advertising the Infiniti Q45 on Monday Night Football, for example, still reaches the biggest number of young and middle-aged males who earn more than $75,000, even though they make up only a tiny proportion of the total audience. More often, however, this approach is money down the drain. Its roots lie more in the cost and compensation mechanisms built into the advertiser/ agency relationship than in genuine cost efficiencies or effectiveness.

Monitoring response

Most companies have monitoring systems that are static - they continue to measure the same indicators over time

Most companies have monitoring systems that are static - they continue to measure the same indicators over time. Action standards are not always adjusted to reflect changes in what the copy is intended to achieve. Maintaining awareness and ad recognition, for instance, may continue to be treated as key benchmarks even when awareness has ceased to be an objective. Usually this happens where the linkage between communication objectives and sales is not clearly understood.

If changing perceptions or reinforcing brand loyalty are key objectives, specially designed market research can be used to measure whether the advertising has achieved the desired effect. In regular tracking studies, however, managers often neglect this option for fear of losing historical comparability. Given sufficient forethought and planning, though, it is possible to satisfy both aims and to change tracking studies gradually so that continuity is maintained without sacrificing relevance.

A more targeted approach

So much for the current state of the game. How should a company set about advertising in the emerging environment? The answer lies, in large part, in redesigning internal processes to be consumer driven and much more targeted, with financial results acting as benchmarks for decision making rather than as the starting point of the planning process. A confluence of factors is driving the trend toward more targeted media strategies: the emergence of new marketing problems, fragmentation among consumer segments and within media, the availability of better measurement and analytic tools, and the desire for greater efficiency and measurability (see the insert "Why target?").

The first step in the transformation to the new approach for measuring advertising spend is to define segments tailored to communication needs. Next, communication objectives must be linked to business goals. Finally, advertising must be valued, component-by-component for each segment (see Exhibit 2). Such a framework is applicable to most products or services, but to illustrate the process let us examine a hypothetical campaign for a luxury sports sedan.

1. Define relevant segments

For advertising purposes, segmentation must focus on grouping customers or consumers together with their media habits. Media segmentation is typically carried out along a single dimension, such as demographics (women of 25 to 44) or behavior (frequent flyers). Combining a variety of additional segmentation bases (such as attitudes and benefits) with the traditional information on demography/geography and entertainment habits (movie going, videotape rental, premium and basic channel viewing) can, however, yield smaller, much more sharply defined segments. These segments could be targeted with specific products such as customized pay-per-view packages offered by local cable companies. For media purposes, such segments need to be split further into groups that share media habits.

In the sports sedan example, one of the key segments is households that currently own a three- to five-year-old Japanese compact sedan which they plan to replace in the next six months, that have a household income exceeding $60,000, and that subscribe to CNN.

2. Develop segment objectives

The next step is to identify the critical attributes of the product or service and their relative importance to the chosen segment. Primary research and such common techniques as regression analysis, or somewhat more complex approaches such as conjoint analysis, are often used. The segment’s current perceptions of the product or service also have to be calibrated. Returning to the sports sedan example, we know from existing market research that perceptions of style, performance, and sportiness are considered equally important by the target group, and that each of these is twice as important as comfort, safety, and durability - factors that would probably be more critical for a family sedan.

Once the critical attributes have been identified, realistic targets - based on past experience and actual product performance - can be set for improving the perception of them. In the example, a reasonable target might be a 10 percent improvement in perceptions of style and a 5 percent improvement in performance and sportiness, two related dimensions.

3. Link objectives to incremental contribution

The relationship between advertising and sales can be highly complex. Each specific case needs to be modeled afresh

Here lies the key to making this approach work. The relationship between advertising and sales can be highly complex, often taking the form of a long causal chain with both risks and opportunities at every link. Each specific case needs to be modeled afresh: both the conceptual model and the parameters that indicate the relationships between links can vary by situation and segment. All the same, a comprehensive monitoring program can capture these relationships on a continuous basis, and research similar to that used for identifying the importance of attributes can be employed initially to build the models.

Again, let us see what this stage entails by looking at the example of the sports sedan. Our targets were improvements in perception of style (10 percent), performance (5 percent), and sportiness (5 percent). Averaged out, the overall target becomes a 6.7 percent change in perceptions (Exhibit 3). The impact of this improvement is a 10 percent increase in the probability that the target segment will consider buying this car.

Since half of those considering buying cars in this category are known actually to visit a dealer (while the remainder may be preempted by a competitor, not live near a convenient dealership, or postpone the purchase), this in turn brings about a 5 percent increase in showroom visits. (Both consideration levels and dealership visits can be monitored periodically at the segment level to verify the validity of this model.)

If 10 percent of dealer visits are converted to actual sales (again, a number that should be readily available with proper monitoring), the result is a 0.5 percent increase in sales to this segment. Multiplying by the size of the segment (500,000) and by the margin per unit ($5,000) yields a net incremental contribution of $12.5 million from advertising the message of style, performance, and sportiness to this segment.

4. Estimate cost to communicate

Since media is often bought for segments larger than a targeted approach requires, this is probably the most difficult step in the sequence. But once mutually exclusive segments have been defined, the total cost of buying a particular media vehicle can be divided among them in proportion to their weight within the wider target group for that vehicle.

In the example, if the total spent on the style/ performance/sportiness message on CNN was $1 million, and this segment represented 20 percent of all of the target segments who watch CNN, then the cost attributed to this segment is $200,000 (see Exhibit 4). Aggregating costs for the campaign over all media yields an estimate of the total cost of communicating this message to this segment - in the example, $10 million.

5. Value the advertising

The final step in the approach is simple: subtract the cost of spending on the segment (as calculated in Exhibit 4) from the incremental contribution (Exhibit 3). The value of advertising the style/performance/sportiness message to the chosen segment is thus $2.5 million. In a real-life context, this calculation can be repeated for each segment/message combination to identify where advertising expenditure will attract the most profitable returns.

Making it work

This kind of system can be put in place in most product categories once the dynamics of consumer behavior are adequately understood

Systems like the one described above can be put in place in most product categories once the dynamics of consumer behavior are adequately understood. A major midwestern bank decided to track the achievement of communications objectives such as awareness of product offerings and convenience features (for example, extended hours). Its target population was prospective customers segmented on the basis of demographics. A study of existing customers suggested, however, that the intended campaign would be unprofitable, since the acquisition effort was too expensive.

Instead, advertising could more productively be directed at either deepening or widening the bank’s relationship with segments of existing customers, determined on the basis of current behavior. With one segment of customers - those who have multiple credit cards or loans with various institutions - a campaign to convince them of the benefits of consolidating their loans with the bank would deepen their relationship, be more profitable, and have measurable impact.

Recent initiatives by GM and AT&T have attempted to achieve the same effect with credit card balances. Banks have long sought the holy grail of "relationship banking," but have usually failed because their segmentation has been done on a demographic basis. When - as in retail banking - you are targeting an existing customer base in a well-defined geographic area, it is relatively easy to monitor media habits, target customers effectively, and measure the impact continuously by monitoring telephone and branch inquiries. If these inquiries fail to convert to new business at the expected rate, that reflects a problem with the product or service, rather than with the advertising.

In consumer packaged goods, the resurgence of retailer-sponsored buying clubs that capture large amounts of information on buying behavior permits behavior-based segmentation at a much finer level than was possible with the limited sample consumer panels sponsored by IRI and Nielsen. The latter do, however, have the ability to link behavior and media habits - an important element in this approach.

For too long advertising has escaped the close scrutiny directed at other areas of expenditure and investment

If this all seems complex and difficult, it is. Successful adoption of the approach requires substantial organizational change, including improvements in analytical skills, the upgrading of customer information databases, and modifications in the planning and monitoring process. All the same, it is overdue. For too long advertising has escaped the close scrutiny directed at other areas of expenditure and investment within companies, such as new plant and equipment.

Things are changing, however. In the automobile market, for instance, J. D. Power and Associates recently announced the launch of a new information system that will track actual transaction prices, sales by model, option configurations, and other information for all marques and models of car at an individual dealer level. Using such a system will give automobile marketers a much better idea of the impact of their advertising on the marketplace.

Easier said than done

Given all the benefits that this new approach brings, the fact that it is not widely employed even in the best of companies is an indication of the barriers to change in the advertising arena. There are a number of underlying causes.

Too many goals

There is a lack of consensus on how advertising works and how its economics or its creativity can be measured

First among the obstacles is the lack of consensus on how advertising works and how its economics or its creativity can be measured. The direct result is multiple (and sometimes conflicting) objectives. Advertising is expected to change perceptions, maintain awareness, reinforce brand loyalty, encourage switching, or influence frequency of purchase. All this is in addition to entertaining, persuading, pleasing, informing, and perhaps even amusing the target customer. Such a broad spectrum of goals makes it difficult to derive a consistent set of objectives that all participants will accept.

Should advertising aim to influence sales in the coming quarter, or does it act as a long-term investment in a brand with a payoff that is years away?

Another area of dispute is the time horizon for measuring effectiveness. Should advertising aim to influence sales in the coming quarter, or does it act as a long-term investment in a brand with a payoff that is years away?

At one durable goods marketing company, the advertising director’s response to the question, "What is the objective of your advertising?" was, "To increase sales." Further probing, however, revealed that the company had no real understanding of the link between advertising and sales. Superior technology and safety were the primary messages it sought to convey in the attempt to attract target consumers. But research showed that its advertising copy actually communicated images of luxury and comfort, reinforcing current perceptions of the product. Little wonder that the advertising failed to have the desired impact - increased sales.

Too many players

Another obstacle to change is the presence of multiple players, both within and outside a company. Most notable is the advertising agency - an influential partner that is traditionally reluctant to support objective measurements of effectiveness for a variety of reasons connected with compensation and control. Even within the company, many players at different levels tend to involve themselves in decision making because of the high profile of advertising. What it says about the company reaches not only the target market but also competitors, intermediate customers, and other stakeholders. Decisions can be hijacked by subjective considerations, to the detriment of research.

Not enough facts

Even when the yardstick is clear, the lack of hard data can still make it hard to measure advertising’s impact

Even when the yardstick is clear, the lack of hard data can still make it hard to measure advertising’s impact. In the automobile industry, for example, one of the prime indicators of advertising effectiveness is the traffic in dealer showrooms. Yet many car companies lack reliable and up-to-date information on the volume and nature of showroom traffic. Dealers may have no incentive to provide accurate information, despite the fact that the industry - including dealers themselves - spends large sums on advertising.

Not enough expertise

Finally, identifying relevant target segments and directing media spending toward them is an activity that requires considerable analytical expertise. Advertisers, agencies, and media alike seem unwilling to invest the effort it would take to make full use of the information available.

Companies should be able to buy advertising pages in only those magazine copies that are mailed to subscribers who meet specific criteria

Advertising in the major national magazines illustrates this issue best. By exploiting a combination of subscriber databases and selective binding technology, companies should be able to buy advertising pages in only those copies that are mailed to subscribers who meet specific criteria on demographics, home ownership, or purchase behavior. Subscribers to Child magazine, for instance, already receive editorial content tailored to the age of the oldest child in the family. Though the same technology can equally well be used to target advertising, the economics that would produce benefits for all parties are not yet in place.

The art of the possible

A company that makes diapers, for instance, could focus its advertising of diapers for infants and toddlers to parents of kids in those age groups, while targeting training pants to parents of children who are a little older, rather than advertise all its products to all of these parents. Newsweek recently introduced a service where readers with special interests - say, in sports - could receive extra coverage in those areas. The experiment stumbled because the magazine asked its readers to pay extra for customization. It might have worked if it had instead asked advertisers to pay a premium for the privilege of reaching audiences with special interests.

Taking the potential a step further, a cigarette manufacturer could choose to advertise only to Time subscribers who are smokers. Using the magazine’s proprietary database, the company could target one message to smokers of its own brands (thus reinforcing their choice), while trying to persuade smokers of competing brands to switch by means of another message supported by promotional vehicles such as coupons. One beverage marketer has successfully adopted this approach and increased its advertising efficiency by 200 percent. It achieved this largely through the elimination of waste - that is, by not advertising to inappropriate consumer categories.

While the capability of selective buying has been available for some time (and has been discussed and publicized heavily in the trade press), magazines have been slow to adopt the technology and persuade advertisers of the merits of targeting because of a natural fear that it will lead to a fragmentation of national advertising pages. Some of this loss is inevitable, however, and magazines may be better off driving the process to their benefit, rather than having their hand forced by advertisers and so losing all control.

Advertisers and agencies have typically seen these new technological possibilities as a way to increase efficiency, rather than as a way to increase effectiveness by tailoring messages to audiences in a much more targeted fashion.

Technology and the desire for greater efficiency will drive marketers inexorably toward more targeted advertising. The very complexity of introducing the necessary mechanisms could act as a source of competitive advantage. Those who follow the new discipline should not only reap better returns on their investment, but also build their information capabilities and improve their understanding of their customers.