Purpose: To measure the audience for an advertisement.

Impressions, exposures, and opportunities-to-see (OTS) are the “atoms” of media planning. Every advertisement released into the world has a fixed number of planned exposures, depending on the number of individuals in its audience. For example, an advertisement that appears on a billboard on the Champs-Élysées in central Paris will have an estimated number of impressions, based on the flow of traffic from visitors and locals. An advertisement is said to reach a certain number of people on a number of occasions or to provide a certain number of impressions or opportunities-to-see. These impressions or opportunities-to-see are a function of the number of people reached and the number of times each such person has an opportunity to see the advertisement.

Methodologies for estimating opportunities-to-see vary by type of media. In magazines, for example, opportunities-to-see do not equal circulation because each copy of the magazine may be read by more than one person. In broadcast media, it is assumed that the quantified audience comprises those individuals available to hear or see an advertisement. In print and outdoor media, an opportunity-to-see might range from a brief glance to a careful consideration. To illustrate this range, imagine that you’re walking down a busy street. How many billboard advertisements catch your eye? You may not realize it, but you’re contributing to the impressions of several advertisements, regardless of whether you ignore them or study them with great interest.

When a campaign involves several types of media, marketers may need to adjust their measures of opportunities-to-see in order to maintain consistency and allow for comparability among the different media.

Gross rating points (GRPs) are related to impressions and opportunities-to-see. They quantify impressions as a percentage of the population reached rather than in absolute numbers of people reached. Target rating points (TRPs) express the same concept but with regard to a more narrowly defined target audience.

Construction

Impressions, Opportunities-to-See (OTS), and Exposures: The number of times a specific advertisement is available to be seen or otherwise exposed to media audiences. This is an estimate of the audience for a media “insertion” (one ad) or campaign. Impressions = OTS = Exposures.

Data Sources, Complications, and Cautions

Data on the estimated audience size (reach) of a media vehicle are typically made available by media sellers. Standard methods also exist for combining data from different media to estimate “net reach” and frequency.

Two different media plans can yield comparable results in terms of costs and total exposures but differ in reach and frequency measures. In other words, one plan can expose a larger audience to an advertising message less often, while the other delivers more exposures to each member of a smaller audience. Table 10.1 provides an example.

* Average frequency is the average number of exposures made to each individual who has received at least one exposure to a given advertisement or campaign. To compare impressions across media, or even within classes of media, one must make a broad assumption: that there is some equivalency between the different types of impressions generated by each media classification. Nonetheless, marketers must still compare the “quality” of impressions delivered by different media.

For example, a billboard along a busy freeway and a subway advertisement can both yield the same number of impressions. Whereas the subway advertisement has a captive audience, members of the billboard audience are generally driving and concentrating on the road. As this example demonstrates, there may be differences in the quality of impressions. To account for these differences, media optimizers apply weightings to different media vehicles. When direct response data are available, they can be used to evaluate the relative effectiveness and efficiency of impression purchases in different media. Otherwise, this weighting might be a matter of judgment. A manager might believe, for example, that an impression generated by a TV commercial is twice as effective as one made by a magazine print advertisement.

Similarly, marketers often find it useful to define audience subgroups and generate separate reach and frequency statistics for each. Marketers might weight subgroups differently, just as they weight impressions delivered through different media differently.¹ This helps in evaluating whether an advertisement reaches its defined customer groups.

When calculating impressions, marketers often encounter an overlap of people who see an advertisement in more than one medium. Later in this text, we will discuss how to account for such overlap and estimate the percentage of people who are exposed to an advertisement multiple times.

Purpose: To compare the costs of advertising campaigns within and across different media.

A typical advertising campaign might try to reach potential consumers in multiple locations and through various media. The Cost per Thousand Impressions (CPM) metric enables marketers to make cost comparisons between these media, both at the planning stage and during reviews of past campaigns. (Technical people like to use mille—from Latin—for “thousand,” hence the M in CPM.)

Marketers calculate CPM by dividing advertising campaign costs by the number of impressions (or opportunities-to-see) that are delivered by each part of the campaign. As the impression counts are generally sizable, marketers customarily work with the CPM impressions. Dividing by 1,000 is an industry standard.

Cost per Thousand Impressions (CPM): The cost of a media campaign relative to its success in generating impressions or opportunities-to-see.

Construction

To calculate CPM, marketers first state the results of a media campaign (gross impressions) in thousands. Second, they divide that result into the relevant media cost:

Cost per Thousand Impressions (CPM) ($)=Advertising Cost ($)Impressions Generated (# in Thousands)

Data Sources, Complications, and Cautions

In an advertising campaign, the full cost of the media purchased can include agency fees and production of creative materials, in addition to the cost of media space or time. Marketers also must have an estimate of the number of impressions expected or delivered in the campaign at an appropriate level of detail. Internet marketers can usually easily access these data (see Chapter 11, “Online, Email, and Mobile Metrics”).

CPM is only a starting point for analysis. Not all impressions are equally valuable. Consequently, it can make good business sense to pay more for impressions from some sources than from others.

In calculating CPM, marketers should also be concerned with their ability to capture the full cost of advertising activity. Cost items typically include the amount paid to a creative agency to develop advertising materials, amounts paid to an organization that sells media, and internal salaries and expenses related to overseeing the advertisement.

Related Metrics and Concepts

Cost per Point (CPP): The cost of an advertising campaign, relative to the rating points delivered. In a manner similar to CPM, CPP measures the cost per rating point for an advertising campaign by dividing the cost of the advertising by the rating points delivered.

Purpose: To separate total impressions into the number of people reached and the average frequency with which those individuals are exposed to advertising.

To clarify the difference between reach and frequency, let’s review what we learned in Section 10.1. When impressions from multiple insertions are combined, the results are often called gross impressions or total exposures. When total impressions are expressed as a percentage of the population, this measure is referred to as gross rating points (GRPs). For example, suppose a media vehicle reaches 12% of the population. That vehicle will have a single-insertion reach of 12 rating points. If a firm advertised in 10 such vehicles, it would achieve 120 GRPs.

Now, let’s look at the composition of these 120 GRPs. Suppose we know that the 10 advertisements had a combined net reach of 40% and an average frequency of 3. Then their gross rating points might be calculated as 40 * 3 = 120 GRPs.

Unless otherwise specified, simple measures of overall audience size (such as GRPs or impressions) do not differentiate between campaigns that expose larger audiences fewer times and those that expose smaller audiences more often. In other words, these metrics do not distinguish between reach and frequency.

Net reach and reach refer to the unduplicated audience of individuals who have been exposed at least once to the advertising in question. Reach can be expressed as either the number of individuals who have seen the advertisement or the percentage of the population that has seen the advertisement.

Reach: The number of people or percentage of population exposed to an advertisement.

Frequency is calculated by dividing gross impressions by reach. Frequency is equal to the average number of exposures received by individuals who have been exposed to at least one impression of the advertising in question. Frequency is calculated only among individuals who have been exposed to this advertising. On this basis:

Total Impressions = Reach * Average Frequency

Average Frequency: The average number of impressions per reached individual.

Media plans can differ in reach and frequency but still generate the same number of total impressions.

Net Reach: This term is used to emphasize the fact that the reach of multiple advertising placements is not calculated through the gross addition of all individuals reached by each of those placements. Occasionally, the word Net is eliminated, and the metric is called simply Reach.

When multiple vehicles are involved in an advertising campaign, marketers need information about the overlap among these vehicles as well as sophisticated mathematical procedures in order to estimate reach and frequency. To illustrate this concept, the following two-vehicle example can be useful. Overlap can be represented by a graphic known as a Venn diagram (see Figure 10.1).

Data Sources, Complications, and Cautions

Although we’ve emphasized the importance of reach and frequency, the impressions metric is typically the easiest of these numbers to establish. Impressions can be aggregated on the basis of data originating from the media vehicles involved in a campaign. To determine net reach and frequency, marketers must know or estimate the overlap between audiences for different media or for the same medium at different times. It is beyond the capability of most marketers to make accurate estimates of reach and frequency without access to proprietary databases and algorithms. Full-service advertising agencies and media buying companies typically offer these services.

Assessing overlap is a major challenge. Although overlap can be estimated by performing customer surveys, it is difficult to do this with precision. Estimates based on managers’ judgment occasionally must suffice.

Purpose: To establish assumptions about the effects of advertising frequency.

Let’s assume that a company has developed a message for an advertising campaign and that its managers feel confident that appropriate media for the campaign have been selected. Now they must decide: How many times should the advertisement be placed? The company wants to buy enough advertising space to ensure that its message is effectively conveyed, but it also wants to ensure that it doesn’t waste money on unnecessary impressions.

To make this decision, a marketer will have to make an assumption about the value of frequency. This is a major consideration: What is the assumed value of repetition in advertising? Frequency response functions help us to think through the value of frequency.

Frequency Response Function: The expected relationship between advertising outcomes (usually in unit sales or dollar revenues) and advertising frequency.

There are a number of possible models for the frequency response functions used in media plans. A selection among these for a particular campaign will depend on the product advertised, the media used, and the judgment of the marketer. Three of the most common models are described next.

Linear Response: The assumption behind a linear response function is that each advertising exposure is equally valuable, regardless of how many other exposures to the same advertising have preceded it.

Learning Curve Response: The learning curve model, or S curve model, rests on the assumption that a consumer’s response to advertising follows a progression: The first few times an advertisement is shown, it does not register with its intended audience. As repetition occurs, the message permeates its audience and becomes more effective as people absorb it. Ultimately, however, this effectiveness declines, and diminishing returns set in. At this stage, marketers believe that individuals who want the information already have it and can’t be influenced further; others simply are not interested.

Threshold Response: The assumption behind this model is that advertising has no effect until its exposure reaches a certain level. At that point, its message becomes fully effective. Beyond that point, further advertising is unnecessary and would be wasted.

These are three common ways to value advertising frequency. Any function that accurately describes the effect of a campaign can be used. Typically, however, only one function will apply to a given situation.

Construction

Frequency response functions are most useful if they can be used to quantify the effects of incremental frequency. To illustrate the construction of the three functions described in this section, we have tabulated several examples.

Tables 10.2 and 10.3 show the assumed incremental effects of each exposure to a certain advertising campaign. Suppose that the advertisement will achieve maximum effect (100%) at eight exposures. By analyzing this effect in the context of various response functions, we can determine when and how quickly it takes hold.

Under a linear response model, each exposure below the saturation point generates one-eighth, or 12.5%, of the overall effect.

The learning curve model is more complex. In this function, the incremental effectiveness of each exposure increases until the fourth exposure and declines thereafter.

Under the threshold response model, there is no effect until the fourth exposure. At that point, however, 100% of the benefit of advertising is immediately realized. Beyond that point, there is no further value to be obtained through incremental advertising. Subsequent exposures are wasted.

The effects of these advertising exposures are tabulated cumulatively in Table 10.3. In this display, maximum attainable effectiveness is achieved when the response to advertising reaches 100%.

We can plot cumulative effectiveness against frequency under each model (see Figure 10.2). The linear function is represented by a simple straight line. The threshold assumption rises steeply at four exposures to reach 100%. The cumulative effects of the learning curve model trace an S-shaped curve.

Frequency Response Function; Linear: Under this function, the cumulative effect of advertising (up to the saturation point) can be viewed as a product of the frequency of exposures and effectiveness per exposure.

Frequency Response Function; Linear (I) = Frequency (#) * Effectiveness per Exposure (I)

Frequency Response Function; Learning Curve: The learning curve function can be charted as a non-linear curve. Its form depends on the circumstances of a particular campaign, including selection of advertising media, target audience, and frequency of exposures.

Frequency Response Function; Threshold: The threshold function can be expressed as a Boolean “if” statement, as follows:

Frequency Response Function; Threshold Value (I) = If [Frequency (#) ≥ Threshold (#), 1, 0]

Stated another way, in a threshold response function, if frequency is greater than or equal to the threshold level of effectiveness, then the advertising campaign is 100% effective. If frequency is less than the threshold, there is no effect.

Data Sources, Complications, and Cautions

A frequency response function can be viewed as the structure of assumptions made by marketers in planning for the effects of an advertising campaign. In making these assumptions, a marketer’s most useful information can be derived from an analysis of the effects of prior ad campaigns. Functions validated with past data, however, are most likely to be accurate if the relevant circumstances (such as media, creative, price, and product) have not significantly changed.

In comparing the three models discussed in this section, the linear response function has the benefit of resting on a simple assumption. It can be unrealistic, however, because it is hard to imagine that every advertising exposure in a campaign will have the same effect.

The learning curve has intuitive appeal. It seems to capture the complexity of life better than a linear model. Under this model, however, challenges arise in defining and predicting an advertisement’s effectiveness. Three questions emerge: At what point does the curve begin to ramp up? How steep is the function? When does it tail off? With considerable research, marketers can make these estimates. Without it, however, there will always be the concern that the learning curve function provides a spurious level of accuracy.

Any implementation of the threshold response function hinges on a firm’s estimate of where the threshold lies. This has important ramifications. If the firm makes a conservative estimate, setting the tipping point at a high number of exposures, it may pay for ineffective and unneeded advertising. If it sets the tipping point too low, it may not buy enough advertising media, and its campaign may fail to achieve the desired effect. In implementation, marketers may find that there is little practical difference between using the threshold model and the more complicated learning curve model.

Related Metrics and Concepts

Wear-in: The frequency required for a given advertisement or campaign to achieve a minimum level of effectiveness.

Wear-out: The frequency at which a given advertisement or campaign begins to lose effectiveness or even yield a negative effect.

Purpose: To assess the extent to which advertising audiences are being reached with sufficient frequency.

Many marketers believe their messages require repetition to “sink in.” Advertisers, like parents and politicians, therefore repeat themselves. But repetition must be monitored for effectiveness. Toward that end, marketers apply the concepts of effective frequency and effective reach. The assumptions behind these concepts run as follows: The first few times people are exposed to an ad, it may have little effect. It is only when more exposures are achieved that the message begins to influence its audience.

With this in mind, in planning and executing a campaign, an advertiser must determine the number of times that a message must be repeated in order to be useful. This number is the effective frequency. In concept, this is identical to the threshold frequency in the threshold response function discussed in Section 10.4. A campaign’s effective frequency depends on many factors, including market circumstances, media used, type of ad, and campaign. As a rule of thumb, however, an estimate of three exposures per purchase cycle is used surprisingly often.

Effective Frequency: The number of times a certain advertisement must be exposed to a particular individual in a given period to produce a desired response.

Effective Reach: The number of people or the percentage of the audience that receives an advertising message with a frequency equal to or greater than the effective frequency.

Construction

Effective reach can be expressed as the number of people who have seen a particular advertisement or the percentage of the population that has been exposed to that advertisement at a frequency greater than or equal to the effective frequency.

Effective Reach (#, %) = Individuals Reached with Frequency Equal to or Greater Than Effective Frequency

Data Sources, Complications, and Cautions

The internet has provided a significant boost to data gathering in this area. Although even online campaigns can’t be totally accurate with regard to the number of advertisements served to each customer, data on this question from online campaigns are far superior to those available in most other media.

Where data can’t be tracked electronically, it’s difficult to know how many times a customer has been in a position to see an advertisement. Under these circumstances, marketers make estimates on the basis of known audience habits and publicly available resources, such as TV ratings.

Although test markets and split-cable experiments can shed light on the effects of advertising frequency, marketers often lack comprehensive, reliable data on this question. In these cases, they must make—and defend—assumptions about the frequency needed for an effective campaign. Even where good historical data are available, media planning should not rely solely on past results because every campaign is different.

Marketers must also bear in mind that effective frequency attempts to quantify the average customer’s response to advertising. In practice, some customers need more information and exposure than others.

Purpose: To evaluate the comparative level of advertising committed to a specific product or brand.

Advertisers want to know whether their messages are breaking through the noise in the commercial environment. Toward that end, share of voice offers one indication of a brand’s advertising strength, relative to the overall market.

There are at least two ways to calculate share of voice. The classic approach is to divide a brand’s advertising dollar spend by the total advertising spend in the marketplace.

Alternatively, share of voice can be based on the brand’s share of GRPs, impressions, effective reach, or similar measures. (See earlier sections in this chapter for more details on basic advertising metrics.)

Construction

Share of Voice: The percentage of advertising in a given market that a specific product or brand enjoys.

Share of Voice (%)=Brand Advertising ($, #)Total Market Advertising ($, #)

Data Sources, Complications, and Cautions

When calculating share of voice, a marketer’s central decision revolves around defining the boundaries of the market. One must ensure that these boundaries are meaningful to the intended customer. If a firm’s objective is to influence internet users, for example, it would not be appropriate to define advertising presence solely in terms of print media. Share of voice can be computed at a company level, but brand- and product-level calculations are also common.

In executing this calculation, a company should be able to measure its total advertising spend fairly easily. Determining the ad spending for the market as a whole can be fraught with difficulty, however. Complete accuracy will probably not be attainable. It is important, however, that marketers take account of the major players in their market. External sources such as annual reports and press clippings can shed light on competitors’ ad spending. Publications such as Leading National Advertisers (LNA) can also provide useful data. These services sell estimates of competitive purchases of media space and time. They generally do not report actual payments for media, however. Instead, costs are estimated on the basis of the time and space purchased and on published “rate cards” that list advertised prices. In using these estimates, marketers must bear in mind that rate cards rarely cite the discounts available in buying media. Without accounting for these discounts, published media spending estimates can be inflated. Marketers are advised to deflate them by the discount rates they themselves receive on advertising.

A final caution: Some marketers might assume that the price of advertising is equal to the value of that advertising. This is not necessarily the case. With this in mind, it can be useful to augment a dollar-based calculation of share of voice with one based on impressions.

Purpose: To understand the responsiveness of demand to advertising.

In Chapter 8, “Pricing Strategy,” we considered the price elasticity of demand, which represents the responsiveness of consumer demand (that is, sales) to a change in price. Advertising elasticity of demand is a similar concept that represents the change in consumer demand from an increase or a decrease in advertising. Price elasticities are almost always negative and less than 1.0, meaning that a given percentage change in price will result in a greater percentage change in sales—in the opposite direction of the price change. Advertising elasticities are almost always positive and less than 1.0; a given percentage increase in advertising spending will cause an increase in sales, but the percentage increase in sales will be less than the percentage increase in advertising.

Armed with an estimate of consumers’ responsiveness to advertising, we can calculate the profit-maximizing amount to spend on advertising. We can do this by comparing the incremental contribution margin resulting from the change in sales revenue caused by the change in advertising spending and compare this incremental contribution margin generated by the advertising to the marginal cost of the advertising.

Construction

The advertising elasticity of demand is simply the change in demand from a change in advertising spending. This can be calculated by fitting an equation estimating advertising response as a function of advertising spending. To do this, you are likely to use historical data, the results of an advertising test, or both.

The formula for advertising elasticity of demand is

Advertising Elasticity of Demand—AED (I)

=Change in Quantity Demanded (%)Change in Spending on Advertising (%)

Unlike price elasticity of demand, advertising elasticity is almost certainly positive as increased spending on advertising should lead to greater demand. Greater elasticity means that demand is more responsive to advertising. When elasticity is higher, spending more on advertising is relatively more advantageous than when elasticity is lower.

Typically, managers fit a constant elasticity model to data. The fitted equation contains an estimate of advertising elasticity [the slope coefficient in the regression of ln(Sales) on ln(Advertising)]. However, when only two points are available, the same formula noted for price elasticity in Chapter 8 can be applied to estimate the point elasticity of demand at current spending levels. This point elasticity can be helpful in determining whether the firm is over- or under-spending, as we shall show.

Advertising Elasticity of Demand,Constant Elasticity (I)=ln( D 2 D 1)ln( A 2 A 1)

In this equation, D₁ is initial demand (sales in dollars), and D₂ is demand after change in advertising. A₁ is the initial level of advertising spending, and A₂ is the level of advertising spending after the change.

Data Sources, Complications, and Cautions

A manager needs to be able to map out (or at least predict) demand at various levels of advertising to use advertising elasticity of demand. This can be challenging for many managers who may not have the right data or be able to run tests to estimate elasticity.

Applying the Dorfman–Steiner theorem also assumes that the variables behave in a predictable fashion. Using advertising elasticity to assess the appropriate advertising-to-sales ratio, it is often necessary to assume that elasticity is constant. This is the sort of assumption that is often easier for academic economists to make than for practicing managers. As such, the “scientific” determination of advertising-to-sales ratios has been less widely used than the apparent power of the technique might suggest.

If the effect of advertising is not smooth, the models are much harder to apply. For instance, say that the impact of advertising spending is lumpy, and you need to reach spending thresholds before advertising has any impact. In such cases, raising advertising spending a little may not increase sales at all. Raising spending by a large amount, however, might, for example, allow a new media channel to be used, thereby substantially increasing sales.

The Dorfman–Steiner theorem result best applies to stable markets with well-established brands. When categories are forming (or undergoing fundamental change), however, brands often spend more than is immediately profitable (even after considering carryover effects) in battling for position (standards wars, for example).

Advertising elasticity and optimal advertising can assist in thinking about advertising effectiveness. The idea is to help managers approach the problem of how much to advertise rather than provide them with an answer that must be applied slavishly.

Purpose: To assess the productivity of advertising in generating additional sales.

The Return on Advertising Spend (ROAS) metric is designed to show the effectiveness of advertising spending. ROAS is not the same as but shares some commonality with the return on investment (ROI) calculations discussed in Chapter 12, “Marketing and Finance.” ROAS uses revenue while ROI uses profit to measure return. ROAS is widely used and typically much easier to implement than the MROI (Marketing Return on Investment) metrics described in Chapter 12. We elaborate on the advantages and disadvantages of ROAS compared to MROI in Chapter 12.

Construction

To calculate ROAS, find the incremental revenue generated from the campaign. The estimate of incremental revenue should be divided by the dollar amount spent on the advertising campaign. When it is possible to identify creative, production, and media buying expenses with specific advertising campaigns, a full evaluation of MROI would also include those expenses in addition to the costs of the media placement.

Return on Advertising Spend (%)=Incremental Revenue Generated by Advertising Campaign ($)Cost of Advertising Campaign ($)

To emphasize that only incremental revenue should be counted, the formula can be rewritten as follows:

Return on Advertising Spend (%)Actual Revenue in Period ($)−=Expected Revenue in Period Without Advertising Campaign ($)Cost of Advertising Campaign ($)

Although ROAS refers to advertising, the same calculation can be used to evaluate any number of marketing tactics, such as promotions, sampling, and additional sales force calls.

Data Sources, Complications, and Cautions

A key advantage of ROAS is that it reflects incremental revenues and advertising costs, two variables for which marketers are typically responsible. Also, estimating ROAS by dividing the estimated incremental sales likely to be achieved by the estimated cost of the advertising campaign being considered is useful to deciding whether to launch a campaign. The very significant disadvantage of ROAS compared to the MROI metric is that the former does not reflect incremental profits, while the latter does.

That said, when price–cost margins are the same (that is, do not vary across campaigns being evaluated), the ROAS metric yields the same rank ordering of campaigns as MROI and is much easier to calculate. If marketers can establish a benchmark minimum ROAS that represents a profitable campaign, the ease of application for ROAS may outweigh the disadvantage of its not reflecting margins. Indeed, in many cases, managers may not have access to the precise margin estimates needed for good MROI calculations.

If a good estimate of margin on sales is available, the minimum ROAS level that is profitable—which we can call the benchmark ROAS—can be calculated as 1/Margin on Sales.

For example, a product with a margin of 50% would require an ROAS of 200%, or $2 of incremental revenue generated for each $1 of the advertising campaign. Similarly, margins of 25% would require an ROAS of 400% to break even, and so on.

Since the ROAS measure is typically estimated for an advertising campaign, it is usually an average ROAS for the total spent on the entire campaign. This is likely to be different from the amount that any additional advertisement would generate.

Finding incremental revenue can be difficult but it is crucial not just to consider total revenue. For more on the challenge of setting a baseline and estimating incremental sales, see Section 9.1.

Sponsorship Metrics

Measuring sponsorship faces all the classic problems of marketing measurement. The benefits gained from a successful sponsorship are diffuse, including many long-term benefits. When 3M sponsors the Super Bowl, the company expects something from it—but not necessarily a massive increase in sales during the next week. When Coca-Cola or Visa sponsors an event, it expects long-term benefits, but measuring these benefits is challenging. This section looks at a number of the ways marketers try to determine the value created by sponsorship.

Purpose: To estimate the value of marketing impressions generated from sponsorship activities.

Equivalent Media Value is used as a simple way to assess the value of sponsorship. It involves directly comparing sponsorship to advertising impressions and asking, “How much would it cost if we had to pay for this many impressions?” What is delivered by the sponsorship is thus assumed to be directly equivalent to advertising exposures, impressions, or opportunities-to-see” (recall, all three are equivalent terms).

Construction

To calculate Equivalent Media Value, you need to estimate of the number of impressions created. This number of impressions is then multiplied by the estimated value of a single impression to come up with the Equivalent Media Value, or the value of impressions created through the sponsorship.

Equivalent Media Value ($)=Number of Impressions Created (#) *Estimated Value per Impression ($)

Equivalent Media Value ($) = Number of Impressions Created (#) * Estimated Value per Impression ($)

The estimated value per impression is often described on a CPM basis (that is, per thousand impressions). When you use a CPM basis, you need to ensure that the number of impressions is also expressed in thousands.

In a single sponsorship, different types of exposures may occur (for example, in a stadium, on television, on the radio). Each type of exposure has a different value per impression. This is a big concern unless you can provide a relatively accurate estimate of the value of each type. If you can provide such an estimate, then it is a simple matter to create a value per type of exposure and total the types to find the Equivalent Media Value of an entire sponsorship. It is important to note that not all exposures are of equal value. An exposure where a consumer is paying attention to the medium is likely more highly valued by marketers than an exposure (or opportunity-to-see) where a logo moves past the viewer’s sightline at speed.

Data Sources, Complications, and Cautions

Clearly, there are a number of assumptions embedded in this metric. For the number of impressions, it is likely to be hard to find a figure that is defensible (see Section 10.2).

Marketers also take a number of approaches to estimating the value of an impression. For example, they might compare each impression to a media source for which there is a more active market (for example, billboard, online, magazine advertising).

Purpose: To consider the cost of a sponsorship compared to the profit generated.

Sponsorship ROI is basically the traditional Return on Investment calculations translated into a sponsorship context.

A key task in assessing a sponsorship might be to divide up the return on a sponsorship into direct returns and indirect returns. Direct returns might include such things as Miller Coors gaining beer “pouring rights” at a sports stadium. This is a valuable concession, and it is relatively easy to estimate the value of such a sponsorship. There are a certain number of ticket holders, a certain percentage of them will buy beer, and these sales have associated costs. To work out the direct return, you simply estimate how much beer, on average, each patron will buy and multiply that figure by the number buying to estimate revenue. There will be significant costs involved in generating this revenue (for example, the product, serving staff, materials). Estimating and subtracting these costs allows you to calculate the margin and the profit.

Indirect returns are much harder to value as they are similar to advertising. They generally generate impressions, so measures such as Equivalent Media Value might be used for indirect returns.

Construction

When used correctly, sponsorship ROI is very much like a standard return on investment calculation (see Section 10.2).

Sponsorship ROI (%)=Incremental Profit from a Sponsorship ($)Cost of Sponsorship ($)

This can be divided into direct and industry returns:

Sponsorship ROI (%)=Incremental Direct Returns ($)+Incremental Indirect Returns ($)Cost of Sponsorship ($)

As with all other ROI calculations, it is important to include only relevant returns and costs. This means if the profit or costs are not attributable to the sponsorship (that is, if they would have occurred regardless of the sponsorship), they should not be included in the calculation.

For sponsorships there are often two types of returns:

• Direct returns, such as the rights to sell the product: For example, a beer sponsor might be awarded the pouring rights in the stadium. This has a value that is relatively easily calculable. It is part of the return to the sponsorship and should not be netted off against the costs.

• Other more indirect returns in the form of benefits such as Equivalent Media Value: These returns can be harder to calculate but are often a very significant part of—or may even be the majority of—the return on the sponsorship. To be a meaningful measure of return, the indirect benefits must be something that the firm values. Specifically, one implicit assumption that justifies the idea that indirect returns are true returns is that the company would have paid for what it gains if the sponsorship didn’t provide it. Equivalent Media Value is valuable if you would have spent that amount on media. The sponsorship saves the media costs, giving a return that is valuable to the firm.

Data Sources, Complications, and Cautions

It can be hard to generate reliable estimates for much of the data in sponsorship ROI. There is no perfect solution to this. One should create consistent measures of, for example, the Equivalent Media Value generated. Ideally, these measures should be agreed to by all parties involved in the sponsorship.

Returns to sponsorship should be added together in the numerator of the equation. Netting off direct returns (such as the profit from pouring rights) from sponsorship costs has the effect of lessening the costs. Generally speaking, this increases the reported ROI, showing an overly positive picture of the success of the sponsorship. This could also cause a problem if the direct profits are greater than the costs (for example, if pouring rights are especially valuable). This would leave the cost of the sponsorship as negative, and the resulting calculation would be confusing.

Caution on the Casual Use of ROI

There are ways that marketers use the term ROI that are not faithful to the financial idea of return on investment. For example, the ANA/MASB survey on sponsorship metrics⁴ saw 55% of marketers assessing the ROI of their sponsorships on sales but this does not include the investment. Those working outside marketing will frequently not recognize measures such as increased sales as being true ROI calculations.⁵ Sales are a perfectly reasonable aim of a sponsorship, but showing increased sales is not the same as showing ROI.

Further Reading

Association of National Advertisers and Marketing Accountability Standards Board. (2018). Improving Sponsorship Accountability Metrics, themasb.org/wp-content/uploads/2018/07/ANA-MASB_Improving-Sponsorship-Accountability-Metrics.pdf.

Bendle, Neil Thomas, and Charan K. Bagga. (2016). “The Metrics That Marketers Muddle,” Sloan Management Review, 3, 73–82.

Dorfman, Robert, and Peter O. Steiner. (1954). “Optimal Advertising and Optimal Quality,” American Economic Review, 44, 826–836.

Farris, Paul W., David Reibstein, and Ervin Shames. (1998). “Advertising Budgeting: A Report from the Field,” American Association of Advertising Agencies.

Forrester, J. W. (1959). “Advertising: A Problem in Industrial Dynamics,” Harvard Business Review, 37(2), 100.

Tellis, G. J., and D. L. Weiss. (1995). “Does TV Advertising Really Affect Sales? The Role of Measures, Models, and Data Aggregation,” Journal of Advertising, 24(3), 1.

Tull, Donald S., Van R. Wood, Dale Duhan, Tom Gillpatrick, Kim R. Robertson, and James G. Helgeson. (1986). “‘Leveraged’ Decision Making in Advertising: The Flat Maximum Principle and Its Implications,” Journal of Marketing Research, 23(1), 25–32.