Welcome to Analysis of ROI for Pharmaceutical Promotion

Finally, Predictable Returns on Promotional Investments
by Milton Liebman

Executive Summary
New computer software and research techniques consider the variables in prescription drug promotion, and determine the best mix for the best bottom line.

The retailer Henry Wanamaker long ago made the famous observation that half of his advertising was being wasted, but he didn't know which half. That may still hold true for dry goods. When it comes to selling prescription drugs, newly developed techniques and high-tech software can tell marketing executives the most efficient use of promotion resources.

The new approach by research organizations has become widely available only within the last year. Research organizations are applying new software for computer analyses of promotional mixes that predict and maximize return on expenditures.

There is a greater focus on sales and promotion today. Some executives claim that bringing the product to market successfully is as important as discovering and developing it. Drug companies' sales force sizes are increasing. Direct-to-consumer advertising of prescription drugs is expected to reach or exceed $1 billion this year. Costly samples are essential to sales.

But, in the business climate of today, marketing management must deliver bottom line results to justify promotional expenditures. To predict and measure return, researchers are now able to put hard numbers on what is considered a soft spend.

The research approach and computer software models differ among firms. Essential to the success of all is the fact that physicians probably are the most highly researched consumer audience. Their professional actions and attitudes have been studied and recorded.

Physicians who are high prescribers in a given category, or are willing to switch brands, or try new drugs early -- these practice traits are recorded in computer databases, along with basic information such as average patient load, hospital privileges, affiliation with managed care organizations, and more.

Based on previous behavior and current practice, the physician profile can be analyzed to reflect expected response to changing conditions for a target audience.

All companies described in this article analyze existing and project changing market information. Using different models and approaches, information is gathered on competing products, ad spend, direct promotional activities, and many other factors.

The market picture is then imposed on the physician profile. How frequently should the doctor be sampled? Will the impact of a dinner meeting exceed its cost? If the ad budget is doubled for this product, how much more market penetration will be achieved? Will it pay to use DTC advertising for this product?

All of the researchers take into consideration what one calls "the complexities of the purchase process for pharmaceuticals," in order to predict and measure return on investment.

Medical Marketing & Media has held discussions with executives and obtained case histories and examples of promotion planning from five top research organizations involved in ROI measurement. The approach that each company uses varies. The goal of each company program is the same -- determine the best way to spend the promotional budget for maximum ROI.

Dispensed Rxs: Measure of ROI
"We assert that until a new prescription is actually dispensed, you have not generated a real return on your investment," says Susan Dietrich, senior consultant to The Plymouth Group, custom research and consulting group of IMS America. "But beyond calculating ROI, the use of intermediate data is extremely valuable for understanding the reasons for effectiveness and for making improvements."

DTC advertising has to have an impact on the chain of events leading to drug dispensing, and the promotional goals of DTC advertising differs for each product, Dietrich says. The chain involves actions by patient, physician, pharmacist and again the patient in terms of benefits from taking the drug (see Figure 1).

The patient must be motivated to visit the physician for the condition and ask for the drug as the first actions in the chain of events, she said. Professional promotion and personal experience, along with DTC awareness, influences the physician's preference for the drug. Considering the patient's condition, the coverage of the managed care plan, and the patient's request, the physician decides whether to prescribe the brand, she continued.

Then, the pharmacist must have the drug in stock and dispense it, rather than substitute another product, Dietrich points out. Finally, the drug has to be well tolerated and improve symptoms for the patient to accept the need to obtain refills.

Measuring along the chain
IMS is the nation's largest supplier of information services to the pharmaceutical industry, including physician profiles and prescribing information. It collects 1.5 billion prescriptions annually, from 135,000 locations, and this year expects to earn $1 billion in revenue.

To measure whether DTC works, The Plymouth Group uses its prescription data-base and can also survey physicians and patients. Pharmacy dispensing and patient refill are monitored (see Figure 2). The marketplace results, promotional costs, and effects of each element provide the data for The Plymouth Group to undertake statistical and financial analysis of promotional spending. The investment can then be justified, or not, and improvements made in the program using the insights obtained.

ROI study for 'Wonder Drug'
The process of calculating the ROI of DTC TV advertising was demonstrated with use of a straightforward example constructed by George Chressanthis, Ph.D., director of marketing science for The Plymouth Group. Because each study is proprietary to the company for which it is done, a simulated product is used to demonstrate the process discussed.

The example involves determination of the ROI for a DTC campaign being tested in four metropolitan statistical areas (MSAs). A TV campaign runs concurrently for six months. Similar MSAs are chosen as control groups. Results are measured in terms of new prescriptions generated.

Each new prescription ideally results in six months of therapy (182 days) if all refills are obtained. In the example, the average selling price that the pharmaceutical company receives per day is $3.00, with a profit margin of 85 percent.

The promotional campaign measured DTC reach and frequency in gross rating points (GRPs) per MSA per month at a unit cost of $100 each. Three months of pretest prescribing data are collected to establish a baseline, and three months of post campaign to establish impact.

To analyze the results, The Plymouth Group addresses a number of statistical questions, such as the type of model, the form of the regression, and the variables to be used. The model is run, diagnostics tests are performed, and comparisons with the control areas are analyzed.

The DTC campaign results
The model results indicated that one GRP produced .23, or less than one-quarter, of a new prescription. In calculating additional profit, one new prescription generates $2.55 ($3.00 x .85) per day, making the profit over 182 therapy days $106.74 ($2.55 x 182 days x .23 Rx) per GRP.

Therefore, there was an increase in ROI of 6.74 percent as the cost of the gross rating point was $100 and the return was $106.74. The total campaign cost $400,000, and it brings additional profit of $26,960.

Dr. Chressanthis says that this example is realistic. GRP cost generally runs $80 for network to $120 for local TV. When evaluating ROI over longer periods, he points out, the cost of capital must be included in the analysis, which at current rates is in the area of 12 percent.

In looking at the ROI, this example illustrates short-term financial results rather than potential increases in long-term brand equity and product loyalty. There are benefits other than short-term ROI, he added, depending on the goals and strategic objectives set for the product.

The new Plymouth DTC-ROI program includes modeling of all promotional activities, such as detailing, journal advertising, and sampling at either a national or sub-national level. A study takes about four to six weeks, once all needed market data and related research is completed. The cost for a custom program is no lower than $50,000, according to Dietrich, and depending on the complexity of the study, can run much higher.

Analyzing a moving market
The R.O.I. Group was formed a year ago for the purpose reflected in its name. It is a division of the research firm Barnes/Hollander, Atlanta, and was started by Steve Barnes, president, and Senior Vice President Dick Anderson, an expert in promotion response analysis. Ty Schuldner joined the R.O.I. Group as vice president, sales and account service.

"An advertising effectiveness model must be rich enough in its analytic structure to encompass both professional and direct-to-consumer advertising, while simultaneously accounting for all other promotional forms," according to Dick Anderson. He finds that the analytic system must start with the three key elements of any advertising campaign, the ad unit, journal schedule, and the product. Anderson emphasized that these elements "must be viewed in the full context of its competitive environment. It should be predictive, rather than retrospective."

There are many reasons why advertising can and does fail, he said, but over half the failed campaigns are due to insufficient financial resources behind the journal schedule. In other cases ad units were simply worn out or never achieved the required attention, recall, or persuasion in the first place.

"What happens to your product is a function of two sources of promotion -- yours and your competitors," he pointed out. "Any effort to measure advertising effectiveness must account for competitive activity and the structure and dynamics of the marketing environment."

The R.O.I. Group considers three key measurement concepts under the product umbrella -- life cycle, market structure, and competitive environment. A host of market and competitive measures interplay. Within the full context of the marketing environment, no promotional element is independent of other promotional elements. "How advertising 'fits' into this environment is neither constant nor trivial," Anderson said.

Analyzed with use of available databases and client information, the R.O.I. Group determines the structure of advertising impact by interacting advertising unit, journal schedule, and the product (see Figure 3). In this analytic process, the group studies the divergence patterns which result from various combinations, patterns, and levels of advertising.

The Group's analytic system and data-base were developed from extensive -- and ongoing -- research on hundreds of advertising campaigns.

Once all data have been integrated, the typical study requires approximately eight weeks to perform. The study investment is a function of the required analysis, according to Anderson. A basic study analyzes a single ad unit, in a single medical specialty, for a specific journal schedule. It statistically blocks, or accounts for, the impact of the other forms of promotion.

A full-blown analysis could study several medical specialties and analyze the separate and interactive roles of professional detailing, advertising, sampling, and mail, as well as DTC print and television campaigns. Depending upon the depth and complexity of the study, the cost can range from $25,000 to $200,000+.

Within its first year, the R.O.I. Group has completed studies for several major pharmaceutical companies.

Optimizing sampling outcome
A goal of researchers is not only to determine, but also to increase ROI on any promotional investment. One large organization is NDC Health Information Services, formerly Source Informatics, purchased six months ago by National Data Corporation. Last year, Source Informatics had $84.9 million in revenue.

The company tracks physician prescriptions and attitudes, and records drug sales, providing a national service. As with other firms, this database combined with information on promotion program information helps determine the success of a campaign.

"We can track prescription drugs on a daily or weekly basis to provide DTC marketers with timely and actionable information," according to Frank Ragona, marketing director of DTC programs. Sales calls, dinner meetings, space advertising, and other promotional activities can be analyzed alone or in combination to determine optimum usage and improve ROI.

Case study on sampling strategy
Patrick Brundage, senior product director for NDC Health's research and consulting group, offered the critical content from an actual project designed to help a client create a more effective sampling strategy. The company and product names have been withheld to protect confidentiality. The case history demonstrates how ROI can be improved without additional expenditure.

The company had six different sample packages to distribute and wanted to determine if they could increase sales and profits if they altered the sampling strategy, retaining its fixed budget. Which doctors should receive samples? How many of each type should they receive?

NDC consultants utilized technology that allowed it to develop the profit function (revenue minus cost) for each of the sample types. The initial analysis showed that only two of the sample types had a significantly positive impact on prescribing. Next the consultants worked with the company to develop "constraints" to incorporate into the model. For example, no prescriber should receive more than 500 samples in 6 months.

Then the software was used to determine the most profitable allocation of the samples for each of the 100,000 physicians in the database, based on historic profiles and prescribing behavior. A projection showed the ROI for the company's historical sample plan, as well as two alternative strategies (see Figure 4).

By changing sampling strategy without increased expenditure, the ROI on both alternative plans could be increased by 60 percent over the historic plan that approached breaking even. More tangibly, the best strategy showed an increase in sales of just under $1.5 million, with a commensurate gain in profit. Impressed with the projections, the company had customized reports prepared that gave the field representatives the recommended level of sampling for every doctor in their territory.

NDC Health needed six weeks to develop the database, construct the promotion response model, and prepare the analysis, after having developed specifications and receiving promotional data from the company. This service is provided on either an annual subscription or an ad-hoc project basis. Annual subscriptions range from $75,000 to $200,000 per market, depending on the level of consultative support provided and the number of markets purchased.

Desktop analysis of promo mix
Analyzing the success of each aspect of a promotion mix and forecasting its impact are the primary answers hoped for from ROI research services. A new software system from Technology Solutions Division of Kelly/Waldron & Co., computer data processing and custom research organization, provides another element -- data compression. This system is used in desktop computers with the client undertaking all of the "what ifs" in real time, according to division President Brian J. Dillon. Instead of selling a service, Dillon says, Kelly/Waldron is selling a management tool.

As with other research programs, a strategic market database is developed for a product, containing sales, marketing, and cost information provided by the company. Commercially available physician profile information is obtained from IMS or NDC Health. Updated information on promotional efforts is added.

Next, software developed by Avi Shatz, president of Intercom, based in Israel, performs predictive and tracking functions. The automated interactive software is provided on lease to the pharmaceutical company, data are updated on schedule, and the client can input subjective and hard data to come up with the best solution, according to Dillon. The desktop unit allows a large number of computer runs quickly, and the results are portrayed graphically for easier use.

The effectiveness of any promotional event or activity can be measured through a software response model of individual physicians exposed, compared to a like control group not exposed, Shatz says. For a seminar, for example, the software can determine the effectiveness of the speakers, the reaction of attendees, and ROI for the event based on additional prescribing volume.

POA analysis improves sales
In its field force application, the software system permits evaluation of alternative Plans of Action (POA) by potential outcome, as predicted by a "response to activity" model. With use of individual physician profiles the optimum number of sales calls per year, per physician in a given territory may be explored.

Selecting the high prescriber for the highest number of sales calls may be the wrong approach, according to software developer Shatz. That doctor may be governed by a restrictive managed care formulary, or may be very heavily detailed, and so forth. A more productive physician, though a lower prescriber, may be a better target, he said.

Determining a POA with the software is graphically shown (see Figure 5). The given territory has 251 physicians. In the targeted promotion plan 102 have been selected for 1,122 visits, or 11 a month during the year based on predictive analysis. The remaining physicians have been grouped to receive fewer calls, say 8, 9, or 10 each. The goal is to obtain greater dollar return per call. This analysis can be applied to each territory in the company, for the greatest impact on prescribing.

Tests of this new system began in January. Four companies are now examining sales calls and sampling for drugs in the cholesterol lowering, antidepressant, oral hypoglycemic, and anti-infective categories, Dillon said. The software system is licensed to a company for an annual fee that may run from $50,000 to $100,000 depending on the number of markets involved. In addition, there is an annual fee of $20,000 per market.

Sales response to promotion
One of the most challenging questions facing today's marketing management is, "What am I getting back for the money I am spending on promotion?" said Marshall Paul, president of PERQ/HCI, the recently merged market research organization, now owned by VNU Marketing Services.

Until now, companies conducting their own promotion research have relied on three distinctly different approaches in measuring ROI, according to Paul. The first finds its genesis in response modeling, a purely quantitative technique that does not evaluate the campaign's quality. A second approach calls for a prospective study with test cells that are matched to control cells. By varying the media mix and spend levels in the test cells, it is possible to achieve greater comfort in the relationship between cause and effect.

The third approach is qualitative. It measures change in awareness and attitude. While this tends to provide a better understanding of why a particular promotion is working, it falls short in providing hard evidence on sales response.

"In the past three years, we have developed an approach which marries the qualitative with the quantitative," Marshall Paul explained, "to produce return on promotion investment (ROPI)." Specifically, a direct mail, photo-illustrated questionnaire measures physician awareness of an ad campaign. Exposure to detailing, journal advertising, other print forms of promotion and sales aids determine sales response to these elements of the promotion mix, Paul said.

The respondent sample of 2,000 potential high prescribers is matched to the respondents' actual prescribing activity. Based on their responses, physicians are split into groups, ranging from those who experience no brand promotion to those who have seen some or all of the media tested. By comparing groups, PERQ/HCI determines the additive effect of various media.

Sales generated due to promotion are determined by the difference in market share and share growth between the promoted group and the non-promoted group. Total sales generated divided by the promotional expenditures allocated to the promoted group are reported as ROPI. Except in unusual situations, $1.25/$1.00 is considered breaking even. The $1.25 figure includes $1.00 to cover promotion costs, $.15 to cover product costs, and $.10 to represent a normal return that could be achieved in the open market.

There is considerable discussion as to which is more important -- how much to promote or what to say. "Because of the qualitative and quantitative aspects of our program and a 12-year history in tracking ads, we can provide strong insights into the issues that cause the results," Paul said. "For example, print promotion usually provides higher message penetration than detailing because print (most notably journals) obtains higher reach into the target audience with greater consistence of message delivery."

Print does not provide a more powerful impact than detailing, in fact the reverse is true. However, print delivers a consistent message to a wider audience.

"The strength of personal selling is the ability to respond to concerns, deliver a persuasive message that is both relevant and believable, provide samples, and use a personal relationship to motivate the physician to use a specific brand," Paul said. "Detailing provides the power behind the marketing effort, while print provides efficiencies."

56 campaigns measured
To date, PERQ/HCI has measured sales response to 56 campaigns, Paul reports (see Figure 6, left). In 41 cases (73 per-cent), the advertising improved detailing's return from $2.28/1.00 to $2.90/1.00 for an average increase of 27 percent. The ROPI improvement is understandable, because the cost per exposure for advertising ($.40) is a fraction of detailing's cost per exposure ($80.00).

In the 15 campaigns that failed to improve detailing's return, we found "problem messages," Paul reported. These were determined to be messages that were perceived as unbelievable, irrelevant, or generic in nature. Accordingly, Paul said, there is strong evidence to suggest that what you say is more important than how much you spend. If nothing else, research documents that greater message quality leads to greater promotion efficiency which, in turn, translates to higher ROPI.

He also found that effective advertising in the absence of detailing normally provides a positive return ($1.97/1.00). "It can be said that journal advertising does double duty -- it brings its own positive ROPI and it adds to detailing's effectiveness," Paul said.

The cost for a research program measuring the effectiveness of a single campaign can range from $16,000 to $64,000 depending on the media being measured and other factors.

"A single ROI calculation is just the beginning," he concluded. "Should the campaign be continued, modified, or changed? If modified, will the next dollar spent lead to an even greater return, or will it erode the current position?

"The tools are now available to help answer these questions and provide direction for optimal spending," Paul concluded. And his statement reflects the goal of all of the ROI programs discussed here.

Reference: Liebman M: Finally, predictable returns on promotional investments. Medical Marketing & Media 1998;33(6):64-74.