Market design, not marketing

Kris Krug // Strategic News Service

Kris Krug // Strategic News Service

By Nick Fritz

Market design has existed in elementary forms since the early 1980’s. Since then, over $200B has been invested in market design technologies. This begs the question: Why haven’t we heard about market design? And why aren’t we any good at it? The answer may be that we are just now attaining the computing power required to really study market dynamics and consumer behavior, according to a panel consisting of R.Preston McAfee, Chief Economist and VP of Microsoft, Michael Schwarz, Chief Scientist for Waze, Google, and Pai-Ling Yin, Associate Professor or Clinical Entrepreneurship at Marshall School of Business, USC.

Market design is to economics what physics is to engineering. It is the set of underlying principles that explains, and allows us to predict, market behavior. Market design seeks to tweak the mechanics of markets in order to best take advantage of market forces and ultimately create efficient markets quickly. A traditional example of market design is the algorithm which matches medical residents with residency positions based on sets of preferences from each group.

The residency matching market design works well because it is relatively simple and occurs in a relatively isolated space with few variables. How do we conduct this same process in more complicated marketplaces? According to the panel, big data holds a clue.

“We are on the verge of a revolutionary period in market design,” McAfee said.

Data from mobile apps is on frontier of market design that is just now being explored. The huge amounts of data coming in from mobile app users may be very useful in determining and predicting consumer behavior.

One of the challenges with this huge amount of data is “triage,” i.e. determining which data is valuable and which isn’t. According to McAfee, there are essentially three types of data:

  1. Data that depreciates in value very quickly, for e.g. day-to-day browser history.
  2. Data that requires large sample sizes to reveal its value, but is valuable over time, for e.g. a consumer searching for ski resort vacations every autumn.
  3. Data that is immediately and durably valuable, for e.g. a professional skier’s enduring interest in skiing products at any time.

Yin thinks that current entrepreneurs in the tech space are best positioned to harness the value of this data because they are able to build in a culture of valuing data from the beginning. One example is a firm called LotaData, which collects vast quantities of data from its customer’s mobile app platforms and aggregates it to reveal consumer trends across communities of users.

However, the panel agrees that valuing data is a difficult problem. How do we value data in a world where some data depreciates in value very quickly and other data is only useful once it reaches a critical scale? This is a problem that has not yet been solved, but the solution will likely drive the future of market design.

To discover more or read other articles from the conference, or our Medium blog.