OPTIMIZING THE MARKETING MIX
by Steve Waldschmidt, modeling analyst, Trajecta Inc.
Today's corporate leaders must direct their companies through an increasingly complex and risky business climate. Critical decisions often must be made very quickly and with incomplete information. The determination of pricing strategies, capital budgeting, sales force distribution, and marketing activity levels all fall within this paradigm. To complicate the decision-maker's problem, budget constraints and manpower limitations must also be carefully considered. How can these leaders make the appropriate trade-offs and drive their companies to success?
The short answer is that these business managers must perform some sort of optimization. Before we go much further, we should define exactly what we mean by optimization in the business world. Optimization is the process of determining an ideal blend of values for all controllable variables. What is a controllable variable? It is a variable that a manager has the power to change. For example, the marketing efforts that are chosen to promote a product (advertising, pricing, warranties, channels, packaging, etc.) are all controllable. There are other variables, such as the demographics of a city in which the product is sold, over which the decision-maker has no control; these variables are not controllable. Thus, the first step in performing an optimization is to identify all of the controllable variables.
Next, in order to choose an optimal blend of controllable variable values, decision-makers must determine how they wish to measure the results of the optimization. In other words, they must identify a goal for the optimization tool to seek. Typically, they choose to either maximize the overall return on the company's investment or to maximize the company's profit. These strategies generally boil down to:
Once the manager has identified the relevant controllable variables and has chosen an output goal, several complicating factors remain. For example, in performing an optimization there may be several combinations of variable values that yield similar returns and profits. In addition, some combinations may be much riskier than others or might not be feasible due to budget or manpower constraints. Determining which solution to choose is the million-plus dollar question. To gain insight into the problem, more and more firms rely on specialized optimization techniques. Finding the best technique for a particular problem depends on a number of factors including the amount and nature of information that is available and budget size.
To get a more concrete look at optimization, let's look at a specific example of its application in today's business world. Source Informatics is a leading provider of prescription, prescriber, and managed care database and information technology solutions for the healthcare industry. Just over a year ago Source Informatics introduced Intellect Optimizer, a software tool designed to optimize promotional budgeting and resource allocation in the pharmaceutical industry. The tool helps drug marketers develop sales call plans that utilize an optimal mix of sales force personnel and promotional items (such as samples) for each doctor engaged.
To take advantage of the underlying trends within Source Informatics' data, Intellect Optimizer applies models developed using neural network based data mining and optimization technology from Trajecta. One advantage of this technology is that it does not create a one-size-fits-all sales call plan, but instead customizes the call plan to target physicians on a doctor-by-doctor basis.
According to Bill Blake, a Senior Consultant at Source Informatics who interacts daily with clients in his role as Intellect Optimizer's technical advisor: "As data collection becomes more and more prevalent, companies will attempt to use their data to determine if their promotions are effective, if they have an appropriate number of sales reps, etc. The problem with some models is that they lack the sophistication to incorporate a company’s total marketing mix into the model. The technology that we are using allows us to do this. Another benefit to this technology is the ability to optimize under constraints.
"Frequently with advanced non-linear models, there are many optimal solutions. Some of these solutions, although mathematically correct, are not realistic business solutions. Our technology lets the user guide the math so that the solutions are mathematically optimal and make good business sense."
By turning the integration of domain knowledge and data into a seamless business process, optimization allows companies to achieve considerable improvements in the management of their resources, resulting in significant cost savings and increased returns. With this technology, complex and risky business problems are being transformed into well-defined solutions every day. These applications are already providing many companies with significant competitive advantages. For this reason, organizations that don't want to be left behind will find it necessary to add optimization tools to their arsenal of corporate weapons.
For more information, see http://www.trajecta.com