Optimizing Value and Cost
I recently came across a New York Times article (“How to Beat the Salad Bar”) that purports to use statistical analysis to offer guidance for improving the everyday life of the column’s readers. Naturally, this opening piqued my interest – data and statistics surround me at work – how can I make them work in making my lunch? Data analysis is often for the benefit of the end users or consumers (you and me) but focused on/funded by the company producing the good or service being researched. Here however, the column’s author and statistician Nate Silver is providing data-driven insight solely for us, his readers.
In the article, Silver proceeds to explain to readers how to get the most bang for their buck at “pay-by-the-pound” salad bars: choose spinach or mesclun over romaine, load on the bacon, go easy on the dressing.
Okay – so Silver doesn’t exactly need to use complex statistical analysis and long feeds of data to come up with these findings, he simply compared shelf prices for the items offered on the salad bar. The article caught my attention for another reason. It made me think about optimizing different configurations of products for varied end-consumers. Silver’s analysis, as pointed out in the article’s Readers’ Comments, does not take into account the value of time saved by not dicing one’s own tomatoes or having to purchase more produce than is necessary for the given meal. As one reader put it, “Cost is not the same as value.” If I don’t like curly greens filling my salad, that optimization (maximizing salad ingredients and cost) using mesclun is useless. Silver’s advice can be followed to the extreme as well – like my colleague who now feels that he has the support to get nothing but bacon next time.
How Your Choice of Greens Relates to Market Research
This is why a tool known as Choice Based Conjoint (CBC), a form of discrete choice analysis, is so valuable in market research. CBC experiments are conducted to more closely resemble actual purchasing decisions. They force participant respondents to make the trade-offs that are part and parcel of the evaluation of different product features at different prices. Respondents choose from a set of alternatives rather than evaluating the alternatives one at a time. CBC can effectively forecast demand for new or improved products and their share of preference among a defined set of competitive offerings at different price points.
Back to the Salad Bar
CBC would be able to effectively tell us the value of, say, beets on the salad bar for different segments of consumers rather than just the cost of beets. Beets may be of high value to a beet lover, but for someone who hates them, they’re a low-value pick. Silver’s article describes a method for grazing a salad bar that optimizes commodity value for cost but not one that optimizes the utility of the salad bar for the consumer who is planning to eat the meal. CBC assures the provider that the right options are available to satisfy the most customers at a price they are willing to pay. When the salad bar is designed with an eye only to profit there is no spinach or mesclun, definitely no bacon, and the dressings are watered down.
Jorie Kirschbaum, Account Executive