Measuring the Effects of Sponsorship
SEI, Inc.

by Bill Harvey, Stu Gray and Gerald Despain

A prior version of this report, and accompanying methodology, was originally presented to the Advertising Research Foundation (ARF) in October 2000 and published in the ARF Journal of Advertising Research in November 2002. This updated version includes additional findings developed through September 10th, 2003, the date of the presentation of a cross-publisher study to the Interactive Advertising Bureau’s Advertiser Forum. Bill Harvey is Senior Vice President and General Manager of OpenTV Research, Co-Chair of the New Media Model Committee of the ARF and is the founder of Next Century Media. Gerald Despain is Chief Technical Officer of Next Century Media. The Sponsorship Effectiveness Index (SEITM) was founded by Next Century Media, American Demographics Magazine, and Studio One Networks to develop a Sponsorship Effectiveness Index that enables advertisers to measure the impact of program sponsorships within various new media environments.

Sponsorship is again on the rise, and marketers have called for new efforts by researchers to measure the business value of this form of marketing communications. The methodologies appropriate to making sponsorship accountable are similar to those used for other forms of advertising, and the effects of sponsorship fit within the continuum of effects known as the ARF Model. What causes Persuasion in the sponsorship context appears, however, to be different from what causes Persuasion in the advertising context.

Background To This Study
In 1999, Next Century Media was approached by Studio One, a major creator and syndicator of sponsored content on the Internet and in broadcast outlets, and was asked to develop a methodology for measuring the effects of Internet sponsorships. In subsequent meetings with industry associations, including the Association of National Advertisers (ANA) and the American Association of Advertising Agencies (AAAA), American Demographics magazine, Studio One Networks and Next Century Media organized SEI LLC., whose purpose is to make those methodologies applicable across all media, including television and event sponsorships. The methodology has been applied first to the Internet and is expanding, using constant yardsticks across all media. Between 2000 and 2003, SEI has executed 28 sponsorship studies for ten top advertisers.

American Demographics, a Primedia Intertec publication, is the exclusive publishing partner in SEI Inc. with special focus on delivering insight into the study's findings. The SEI findings in this paper are presented by kind permission of American Demographics Magazine.

Current Questions
Sponsorships were the original advertising form on radio and television. Sponsored events today generate more total dollars than all media advertising combined (from a combination of sponsor expenditure plus ticket sales). On the Internet, sponsorship revenues are expanding faster than banner revenues. In the future, the rise in penetration of Personal Video Recorders such as TiVo, as well as the duplication of PVR function by cable and satellite set top boxes is expected to increase the consumer’s ability to avoid normal advertising. Pundits are proclaiming that the future shall be a new Golden Age for sponsorship, product placement, and cast presenter commercials, in an effort by marketers to keep up the consumer’s diet of marketing messages in face of the new hurdles – avoiding zapping by merging into the programming.

Despite all of the indicators pointing to the need for a clear understanding of the business value of sponsorships, marketers today remain unsure of how sponsorship works and how to properly measure its business value. Marketers see sponsorship as something different from advertising – but there has been no general clarification of how sponsorship differs from advertising, and what this implies in terms of how to make sponsorship accountable. Thus there has been a recent groundswell of interest expressed by advertisers and agencies alike in devising new forms of research to measure the business effects of sponsorships in all media – including events.

Previous Studies
In 1968, we were privileged to work on a C.E. Hooper project for a pet food advertiser to measure the effects of a sponsored Special on CBS in prime time. The cast presenter commercials were seamlessly integrated into the program. Hooper was measuring the scatter plan for that advertiser as well, and applied the same on air testing methodology to measuring the Special. The attitude shift and commercial recall measures for the Special were approximately three times the scatter plan averages. In the latter case, what was being measured was a combination of product-sell commercials plus sponsorship. One indeed wonders what the recall and attitude shift results would have been if the sponsorship did not include product sell commercials. Today, sponsorship generally does not include product sell – although perhaps it might in the future if the combination turned out to be as potent in general as in the case just cited. This potential potency suggests that the industry develop its sponsorship measurement tools so as to be able to distinguish causes and effects to this degree – i.e. separating out the effects of sponsorship itself from accompanying product sell advertising, if any.

In the same year, we were also fortunate in participating in another Hooper study, which used approximately 250,000 coincidental calls to measure the immediate recall of TV commercials from 10AM to 10PM in each local time zone. By 1968 there were only a handful of sponsored programs left on television, as the scatter plan had become the dominant TV media strategy. The immediate commercial recall on these sponsored programs was about 2.5 times the average scatter plan level.

In 1990, CBS commissioned Norman Hecht Research to conduct a series of studies measuring the effects of five prime time TV Specials. The advertisers were Anheuser-Busch,AT&T, Chrysler, GM, Hallmark, McDonald’s, Pepsico, Reebok, and Valvoline. The Top Of Mind Brand Awareness was increased up to +230% with a median increase of +17%. Top Of Mind Advertising Awareness was increased up to +500% with a median increase of +24%. The most significant measure, at the persuasion level, was a Purchase Intent increase of up to +40% with a median increase of +20%. As a benchmark, at about the same time, one of the leading commercial pretesting companies, rsc, found that persuasion increase caused by the average TV commercial tested was on the order of magnitude of +5%.

There are other historical cases as well, suggesting the higher efficacy for sponsorship versus straight advertising. We continue being interested in exploring this further and in better understanding what makes sponsorship work

Hypotheses Of The Sponsorship Effectiveness Index Studies

1. Sponsorship has recall and persuasion effects like traditional advertising.
2. Sponsorship operates through different cognitive processes than those utilized by traditional advertising.
3. Advertising changes the consumer’s perception of a specific product while sponsorship changes the consumer’s perception of a specific sponsor – which can rub off positively on the brands of that sponsor in terms of willingness to purchase those brands.
4. Sponsorship positive business effects are maximized when:
   a. The target audience is highly involved in the subject of the program or event being sponsored.
   b. That audience perceives there to be a scarcity of content in the subject area relative to their appetite for it.
   c. The program/event is executed so as to produce a high level of satisfaction with the material by the audience.

Methodology
When Internet users click on a link to a sponsored program they are intercepted by an invitation screen.
Before they are admitted to the program, they must answer yes or no to a question as to their willingness to answer a few questions “to help us improve the program” which they will be asked upon leaving the program. They may be offered an incentive ranging from a free magazine to a $15 gift certificate at a popular book chain, depending on the degree of the advertiser’s willingness to incent respondents.

Users answering “no” are allowed to go to the program and are not contacted again. Users answering “yes” are also allowed to go to the program, and a random half of them are sent to a version of the program where the sponsor is not mentioned (the control group). Everything else about the program is identical across the exposed group and the control group. As respondents leave the program by any means (e.g. click on a banner, hit the back button, etc.) they receive a short onscreen questionnaire, which is the same for both the exposed group and control group. A person's status (agreed to participate or not, control or exposed group, already responded to questionnaire or not) persists across multiple sessions to prevent duplication of participants or showing sponsorship logos to the control group.

The questionnaire begins with questions about the editorial content and how it might be improved. It then goes on to questions relevant to measuring sponsorship effectiveness, which are adapted for each sponsor’s studies to reflect communications measures that advertiser has found to be most predictive of sales, and which are therefore routinely used to assess advertising performance in tracking studies and in commercial pre-testing. The difference in these measures is attributed to the effect of sponsorship, as all other variables have been held constant.

SEI's software can be calibrated so that all users clicking to go to a sponsored program may be intercepted, or a specific percentage of them can be intercepted.

Findings

Consideration Set
The key success measure for many advertisers is Consideration Set. The consideration set is that group of brands within a product category that the particular consumer is willing to buy. The brand decision at each purchase occasion balances mood, situation, availability, price/promotion, and advertising recency – among the known factors – to select one brand (or more) out of the present consideration set. To be in this set gives a brand a chance of being purchased; to be out of this set is to have a virtually zero chance of being purchased.

Automotive marketers have the following footrule: in buying a new car, the average consumer considers six makes, test drives three, and buys one. This points up the importance of being in the consideration set level of the car buying funnel.

Table 1 shows the combined SEI studies across four different packaged goods brands and four different automotive brands. The comparison is on willing to consider between those who saw the sponsor on the program versus the control group who saw the same program but without the sponsor being mentioned.

The average lift for packaged goods was 25% (one quarter again as many consumers). The average lift for automotive was 55% (half again as many consumers).

Table 1 Sponsorship Effect on Consideration Set for a Car Maker

Purchase Intent for Packaged Goods
In SEI studies for four packaged goods products we asked about purchase intent, which across these studies increased by an average of +61%.

Table 2 Sponsorship Effect of Intent-To-Purchase

Brand Perception
A major clue to the way that sponsorship works is provided by a third measure, Brand Perception. The underpinning of today’s Persuasion measurement is a cognitive attitude shift model in which (a) attitude is a predisposition to behavior and therefore predictive of behavior (b) the predisposition to buy a brand is the mathematical resultant of (b1) the desirability weights that a specific consumer places on various benefit dimensions of a specific product category and (b2) the degree to which that consumer perceives each brand in the category in terms of its ability to deliver each of those benefits.

Horace Schwerin, Al Achenbaum, Russ Haley, Jack Landis and many others deserve credit for establishing this important model (not the only important model) of how advertising works. The present work is aimed at helping sponsorship catch up with advertising in terms of having a similar level of understanding of how the process works.

Within the latter context, in the sponsor’s product category, having one’s brands perceived as one of the best in terms of product attributes would be a good thing; therefore, we would expect that if sponsorship worked like advertising, then given the positive results seen with the other data above, we would expect that the perception of the sponsor’s brands as one of the best would be higher in the exposed group than in the control group. For some studies, we asked the respondents to rate their attitude towards the sponsoring brand. We did not, however, see the expected consistent and large lift (commensurate with the consideration/intent measures) in product attribute perception among those exposed to the sponsorship.