We are a specialty consulting firm exclusively focused on measuring and improving the financial return from marketing investments. With experience in dozens of industries, we use a broad toolkit of unique approaches to find just the right way to break through the political, cultural, and structural obstacles to help you crack your toughest measurement challenges.

View our two new webcast series, designed to help you meet the challenges of marketing measurement and resource allocation.

Classic Marketer Seeking Stimuli That Prompt Profits

Moving a dime-store standard to the Web has its challenges.

Children, parents, and educators expect to see the yellow-and-green Crayola logo on a huge array of art supplies when they walk down the aisles of small-town five-and-dimes or big-box mass merchandisers, even grocery stores. But Crayola wanted them to buy crayons, modeling clay, and construction paper online, too.

So when the brand debuted online, it employed experimental design to identify the attributes of an e-mail that would drive customers to Crayola.com and induce them to purchase. Crayola hit upon a winning subject line, salutation, call to action, and promotion with which to entice parents and teachers to use the new e-commerce site and art-project portal.

It is difficult to isolate what drives consumer behavior in today's market environment, thanks to the proliferation of media and messages. And consumer behavior often does not originate with one product or promotion attribute, unfortunately, but with combinations of stimuli, of which there can be literally millions. Experimental design (also called multivariable testing) quantifies the effects of independent stimuli on behavioral responses and therefore can help marketing executives analyze how the various components of a campaign influence consumer behavior.

In a Harvard Business Review article, "Boost Your Marketing ROI with Experimental Design," Eric Almquist and Gordon Wyner, vice presidents at Mercer Management Consulting in Boston, do a remarkable job of explaining the experimental design process, comparing it to the common-test-and-control-cell method, and illustrating it with a real-life project conducted by Crayola.

The science of experimental design lets people project the impact of many stimuli by testing just a few. By using mathematical formulas to select and test a subset of attribute combos representative of the complexity of all of the original variables, marketers can model thousands of stimuli accurately and efficiently for variable cause and effect, not simply correlation with customer behavior.

The authors note that JPMorgan Chase, Household Finance, Capital One, and America Online — companies with large numbers of customers that face rapid and constant change in their markets and product offers — have applied some form of experimental design to their marketing.

How does it work, you may ask, to not test every permutation but still get specific recommendations for attribute changes and adaptation? Using fractional factorial design, marketers select a subset (fraction) of combinations to test. "Factorial" refers to the crosses of individual attributes. Using logistic regression analysis, the marketing measurers can extrapolate from the results of the experiment the probable responses to all permutations.

Different types of experimental designs can be used when the experimental objectives vary. Screening designs, for instance, can efficiently test very large numbers of attributes towards the selection of a small number to test more thoroughly.

Almquist and Wyner make experimental design nearly a no-brainer move toward smarter marketing by explaining in layman's terms the basics of the process and illustrating them with simple graphics.

To read the entire article, "Boost Your Marketing ROI with Experimental Design" by Eric Almquist and Gordon Wyner, from the October 2001 Harvard Business Review, visit http://harvardbusinessonline.hbsp.harvard.edu.