Monitoring the health of prescription drug advertising can be challenging for a variety of reasons. Finding the target audience — doctors with sometimes rare specialties — can be expensive and time-consuming. The message content of the advertising can be complex, often requiring multiple pages of explanation and proofs of performance. And powerful research diagnostic instruments, which are essential for pinpointing areas of an ad that require creative treatment, are critical. For all these reasons, testing multi-page drug advertising on the Internet would appear to be contraindicated.
For a number of doctor-targeted ad tests conducted in the past few months, recruiting doctors for an on-line interview was found to be significantly faster and cheaper than conventional off-line tests, while producing comparable insights and conclusions. Ad page length, running up to six pages, was found to be correlated with on-line measures of break-through power in an intuitive way, with a high degree of face validity, despite the fact that the Internet removes the tactile experience of holding a paper journal in your hands. And a clever way has been found to replicate on-line the diagnostic insights which come from tracking the reader’s visual path through the ad via eye-tracking.
It’s the state of the art for the advertising research practitioner.
Reduced Cost and Timing
Doctor interviews are among the most expensive and time consuming for advertising researchers to collect. High incentives are needed to motivate doctors to not only take the time to do the research interview but also to travel to a central location where the interviews are collected. This can be quite a problem when you are looking for specialists who may be spread out geographically, perhaps far from advertising research facilities that have specialized equipment for eye-tracking. This factor adds to the time as well as the money needed to collect these interviews.
The following graph demonstrates the potential advantage, in terms of cost and time in field, for on-line doctor interviews compared to those conducted off-line. This is based on five ad tests among doctors for which parallel off-line tests were conducted. (See Figure 1 on the next page.)
The bottom line is that on-line testing among doctors can reduce research costs by nearly forty percent and reduce the turnaround time for a study by almost one third when compared to similar research designs for off-line research.
Incidentally, previous “research-on-research” work by Ameritest (reported in Young, 2002) showed that results from on-line print ad tests are very similar to those conducted off-line. Actually touching the paper the ad has been printed on does not seem to be important. It’s the creative idea that matters!
The dosage of information in a doctor-targeted ad is usually much higher than it is for consumer advertising. It usually contains highly technical language describing specific product claims and support points, graphical demonstrations of quantitative clinical results, visuals dramatizing the emotional benefits of successful patient treatment, as well as therapeutic qualifiers and legal disclaimers. Frequently, multiple pages of advertising space are needed to carry all this content.
From an advertising research standpoint, one key measurement issue is to understand the value of an expensive multi-page media buy. In theory, we would expect that purchasing more pages would lead to higher levels of attention among readers. In practice, that is exactly what we found. (See Figure 2 on the next page.)
The Stopping Power of the advertising is consistently related to the number of pages in the ad, with single-page ads getting the lowest scores on average and the ads with the greatest number of pages getting the highest scores.
Importantly, this does not mean that the creative strength of the execution doesn’t matter. Quite the contrary! When we control for the number of pages, we find substantial discrimination between the performance of individual executions. It’s just that page length is one of the variables you have to take into account when interpreting advertising test scores.
How do we measure “Stopping Power” on the Internet? It’s simply a measure of the attention-getting power of an ad when it has to compete for the reader’s attention against other ads in a clutter portfolio. It is measured by giving respondents a portfolio of print ads to look through. In the Internet version, the test ads are shown electronically, with the respondent having the freedom to click forward and backward through the portfolio and spend as much time on each ad as desired. After looking at all the ads, respondents are asked to recall all the ads they found to be “interesting.” Respondents recalling the test ad with interest, as a percentage of the total number of respondents interviewed, are counted toward the Stopping Power Score.
A Practitioner’s Heuristic Model
Stopping power is, of course, only one measure of advertising performance. Other measures may also be important, such as how well branded the ad is, what it communicates, how “sticky” it is in terms of holding reader attention over time, how much emotion it generates, how likeable it is and whether it motivates the reader. We believe that all these measures are relevant to understanding the complex, multidimensional aspects of advertising effectiveness. Arguing for one measure over another frequently sounds like the fable of the three blind men arguing about the elephant. So, in the on-line interview we collect them all.
The problem for modern researchers is one of synthesis. How do you keep track of, and integrate your thinking about, these different views of the advertising? As shown in Figure 3 (on the next page), Ameritest has constructed a heuristic model — a teaching model — to provide a roadmap for interpreting all that data. The interested reader is referred to Young (2001) for a complete description of this approach and to Young and Cohen (2004) for an account of how key measures in this model were validated against the judgment of seasoned creative directors.
Here’s a brief tour of the model. In the model, information is arranged in a hierarchy that bridges the divide separating report card systems and diagnostic systems.
At the top is what pre-testing is supposed to predict: in-market results. One level down are the evaluative measures that provide the report card portion of the analysis. Two levels down are the diagnostic measures that are correlated with, and therefore explain, the evaluative measures above. The arrows in the model highlight the primary relationships or correlations between the different variables.
Essentially, the model says that for any print ad to be effective it must accomplish three things:
1) It must get noticed and attract a reader.
2) The reader must know the identity of the brand sending the advertising message.
3) Once the ad has the reader’s attention, it must motivate — e.g., generate an interest in prescribing the drug.
Other variables in the model are important only insofar as they help to explain the variables of Stopping Power, Brand Linkage and Motivation — they are Diagnostics. For example, stickiness is not important in and of itself, but it is a key component of the flow of a reader’s attention through the ideas and images in the ad.
A Powerful Diagnostic Instrument
For off-line print testing, the “standard” of diagnostic advertising research is eye-tracking. Unfortunately, this is not an option for on-line researchers. Yet if we understand why eye-tracking information is so useful as a diagnostic, we invent other methodologies for approximating this information in an on-line environment. Our approach is to use a patented, Internet-age version of the classic tachistoscope (or t-scope) methodology. It works because inside every computer ever built is an extremely accurate clock—it is fundamental to the way computer processors work.
For the “eye-tracking” part of the interview, the ad is shown to the respondents for brief, controlled periods of time. The first exposure is one-half second (just long enough for most respondents to see only one thing in the ad); the second exposure is for one second; and the third exposure is for four seconds. After each exposure, respondents are asked:
• What did you see?
• Where exactly in the ad were you looking? (This is recorded on a response grid.)
• If this ad was in a normal magazine, how likely would you be to continue looking at or reading the ad?
The answers to these questions provide the following insights: the ad’s entry point or hook (which is key to understanding Stopping Power), the order in which the respondent “shops” the information in the ad (which is key to Communication), the stickiness or holding power of the ad, and how long it takes respondents to register the brand identity. Figure 4 (below) shows an example of the path of a reader through the page of a print ad.
If necessity is the mother of invention, then serendipity may be the father of scientific experiment. With experience we have learned that this approach actually has one advantage over eye-tracking. While the output of this process is something that looks very much like an eye-tracking map, we are not measuring behavior, not simply where the eyeball is pointing, but rather we are measuring perception — the movement of the mind through the ad. With verbatim comments about what the consumer was thinking while they were looking at a particular part of an ad, we gain additional insights — e.g., how quickly a respondent recognizes a celebrity, or what message is taken away in the first second of glancing at an ad.
Finally, additional tools can be built into the on-line interview to make this channel of information gathering more effective. For example, for copy-intensive parts of an ad, a rollover “magnifying glass” can be used to make it easy for readers to study even the legal disclaimers in the ad — if they are so motivated.
The prognosis looks good for the future of on-line print testing of pharmaceutical advertising. Internet testing generates the performance measures that make intuitive sense — and line up with off-line results, as we have reported in other studies. It also provides the researcher with powerful diagnostic measures that explain these performance measures and help contribute to the process of optimizing advertising executions. Finally, by substantially improving the cost/value equation, it is more practical than ever before for quantitative researchers to play a valuable role in the birth of healthy advertising.
Young, Chuck E. “Researcher as Teacher: A Heuristic Model for Pre-Testing TV Commercials.” Quirk’s Marketing Research Review, (March, 2001).
Young, Charles. “High-Touch vs. High-Tech: Can You Test Print Ads on the Internet?”, Quirk’s Marketing Research Review, (March, 2002).
Young, Charles, and Cohen, Larry. “Creative Awards vs. Copytesting,” Quirk’s Marketing Research Review, (April, 2004).