Home   |   Login   |   Register      

 

    

 

Working Paper

Local Neighborhoods as Early Predictors of Innovation Adoption

Jacob Goldenberg, Sangman Han, Donald Lehmann, Janghyuk Lee, and Kyung Young Ohk, 2010 [10-104]

View/Order Item >>
Comments from members >>

Managers spend considerable effort trying to find ways to predict the behavior of large populations, particularly in marketing new products. In this report, Goldenberg, Han, Lehmann, Lee, and Ohk examine the role of predictive neighborhoods (groups of connected individuals whose adoption patterns evolve similarly to, yet earlier than, overall population behavior) in forecasting adoption of new products.

With data drawn from Cyworld.com, a social network website in Korea, they create a network of 114 neighborhoods and track the diffusion process over the network of new items introduced by CyWorld to users.

They begin by examining individual adoption within neighborhoods, specifically the influence of local characteristics on the timing and speed of adoptions. They find that an individual’s local neighborhood has a significant influence on his or her adoption behavior.

Individuals who receive information from a greater number of sources (greater in-strength) become exposed to information, and hence tend to adopt, earlier. Individuals who share information with a greater number of sources (greater out-strength) also tend to adopt more products, consistent with the opinion leadership literature.

Consistent with findings that hubs increase the speed of adoption, their study shows that density—how tightly connected members of a neighborhood are—affects the speed of adoption.

The betweenness of the neighborhood (that is, how closely linked it is to other parts of the network) increases the number of adoptions. This is consistent with the neighborhood’s role as a broker of information to different neighborhoods, increasing the likelihood of exposure to products and consequently increasing the number of adoptions.

In a second analysis, the authors examine whether some neighborhoods can be reliably used at an early stage of the product introduction process to predict overall network adoption behavior.

Using the same set of 114 neighborhoods from the Cyworld.com data, they identify the top 10 items adopted (the “mega hits”). To determine how well particular neighborhoods predicted adoption in the total network, they used the correlation between (1) adoption of the mega-hit in a neighborhood at the time that 5%, 16%, and 50% of the eventual market had adopted it and (2) eventual total network adoptions. They then examined which characteristics were associated with the neighborhoods that were the best predictors of overall adoption.

They found that large and central neighborhoods that adopted early were the best predictors and their predictions were more accurate than those of random sample of the same size. By contrast, small dense neighborhoods were relatively poor predictors of eventual adoption, possibly because of their closed and isolated nature.

Overall their study demonstrates the potential of using identifiable clusters (predictive neighborhoods) to improve the accuracy of new product adoption predictions within a market. While the study focused on post-launch predictions, firms might use predictive neighborhoods as the basis for pre-launch tests as well.

Jacob Goldenberg is Professor of Marketing, School of Business Administration, The Hebrew University of Jerusalem and a member in the research center at Ono. Sangman Han is Professor of Marketing at Sungkyunkwan University, Korea. Donald Lehmann is George E. Warren Professor of Business, Graduate School of Business, Columbia University. Janghyuk Lee is Associate Professor of Marketing, Korea University Business School, Seoul, Korea, and Kyung Young Ohk is a postdoctoral fellow at the Massachusetts Institute of Technology.

Acknowledgments

The authors would like to thank Dominique Hanssens, Scott Neslin, Vithala Rao, Danny Shapira, and Olivier Toubia for their valuable comments.

---

Comments from members

Please login to view and/or submit comments

---

View/Order Item

Online version
Your Price: $18.00
Hard copy
List Price: not available Your Price: not available