Web Recommendation Engines Can Miss The “Long Tail”
The ability for eCommerce sites to drive sales of the “Long Tail” has been well-publicized since Wired editor Chris Anderson’s article and book of the same name. The theory states that businesses with strong distribution power can sell a greater number of hard-to-find products at small volumes than of large volumes of mass-market products. Yet the recommendation engines employed by these same eCommerce sites may suppress the diversity of purchases, reducing the Long Tail effect.
That’s the hypothesis stated in a new paper by Daniel M. Fleder and Kartik Kosenagar, both of the University of Pennsylvania’s Wharton School of Business.
The paper, Blockbuster Culture’s Next Rise or Fall: The Impact of Recommender Systems on Sales Diversity, examines two alternative views of recommendation engines: one, that they help consumers discover new products, therefore increasing sales diversity; and another, that recommenders simply reinforce the popularity of products which are already popular.
The authors reach four conclusions:
- Some common recommenders reduce sales diversity, as they recommend based upon sales volume and ratings.
- In accordance with the above, it is possible for individual-level product diversity to increase, while aggregate diversity decreases.
- Recommenders intensify the effects of chance events on market outcomes; in other words, marketers can create blockbusters by early, high sales, which could be chance events.
- Design of the recommendation engine can drive outcomes; an eCommerce site can be designed to drive the Long Tail or to further drive sales of existing blockbusters
There’s little doubt that recommendation engines drive additional sales for eCommerce businesses. While individual consumers perceive that recommendation engines increase the diversity of products they are exposed to, this paper makes clear that on the aggregate level, design of the recommender system plays a large role in whether that’s actually the case.
The paper, Blockbuster Culture’s Next Rise or Fall: The Impact of Recommender Systems on Sales Diversity, is available for free download. A podcast is also available.
You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.