The Science of Demand
In “The Science of Demand” (MM March/April 2004), Almquist, Kon and Bock provide a great overview of the application of discrete choice models to the problem of estimating demand for new products. These types of models have been in wide use in marketing applications for more than a decade, and have established a strong reputation and following as tools for product optimization, pricing work, and demand estimation. The authors also correctly point out that the discipline of creating reasonably accurate pre-market sales forecasts is key to reducing failure rates for new products and improving companies' ROI on new product development.
Where the article may lead the reader astray, however, is in its depiction of the model – and especially the consumer response aspect of the model – as the star. The real value of this type of work comes from companies doing the rigorous homework required, of describing their assumptions about the market, and developing marketing plan for the new product launch.
Over many years of testing new products, I was constantly struck by how many otherwise worthy new products died before going to market, or after they were on the shelf, simply because the companies developing them were unwilling to spend enough on marketing to support them, did not achieve sufficient levels of distribution, or pulled marketing support before the product was established.
While Almquist et al. mention in passing the need to “begin to estimate the marketing effort that will be needed to increase awareness, familiarity, and take-up”, they give short shrift to the importance of understanding the impact of marketing support in translating consumer demand into new product sales. The impact of the company's marketing plan on sales is at least as great as the impact of the level of consumer interest in the product. And while there is usually no amount of money that can make a bad product idea fly, an inadequate or poorly thought-out marketing plan can kill even a great new product idea.
Fortunately there are models that can estimate the level of new brand awareness a specific marketing plan will achieve, and that can quantify the impact of a channel distribution plan on sales. In the same way that choice models estimate the sales impact pricing, branding and feature bundles, these marketing reach models can estimate the effect of changes in budgets, and in allocations of resources to alternative marketing activities, as well as the timing of those activities.
Running these types of simulations not only gives a richer, more accurate picture of a new product's potential sales performance, over time these simulations make companies smarter about how to execute the launch of a new product. Coupling these marketing reach models with the power of choice modeling as described in the article is really what “The Science of Demand” is all about.