Is There A Business Case for Click Stream (TH1G)

From IIW

Session Topic: Is There a Business Case for Click Stream (TH1G)

Convener: Sid Sidner

Notes-taker(s): Rick Campbell

Tags for the session - technology discussed/ideas considered:

Discussion notes, key understandings, outstanding questions, observations, and, if appropriate to this discussion: action items, next steps:

Notes located:

Overview: Is there a business case for something similar to the Credit Bureau, where multiple retailers send in their click-streams and then pay to get back an aggregated version of the click-stream, but with the individual controlling access to the aggregated data.

How do you get started?

Amazon, e. g., has a lot of click-stream data, but not necessarily a lot of incentive to “give away” their data.

Little guys (Think Geek, etc) have more incentive, and if things were started up, it could reach a critical mass that might make the big players

“Nobody is going to get sued for collecting their own data” (?)

(aside) Someone has a system for adding annotations to arbitrary places on the web. There is some legal argument that if you have an overlay, that might be a derivative of a copyrighted work, but if you did “side wiki”, you might avoid that. Some else noted that there is case law to suggest that it depends on where it is done. If the “mash up” is done on your server, you have legal exposure. However, if you have it mashed up on the client, at the request and with the consent of the end user, you have no exposure. Ad Block has established (?) that the user may add things to their client that are at odds with the owners of the sites that the user is viewing.

Another note: These ideas are great, but there may be a need for some form of legal defense fund, but the existing case law may mitigate the danger.

In addition to considering the motivation/incentive of the retailers, there must also be incentive for the end-user to use the system. Opportunities exist:

  • You could enable cross-site features that are currently limited ot single sites, e. g. when Amazon says, “You purchased this item on XXX” or Google saying “You visited this site on XXX”
  • How about a message like this: You’re looking at a desired product on Walmart and get a message saying that the item that you want is available for less at Target. (this can incentivize merchants as well.)
  • Entering into the market might not be an explicit event -- over time, you might build interest in something. By the time the user realizes that they’re actively looking for a product, they may already

It might not be too difficult to collect click-stream data.

VRM - You might create a search engine for retailers to search for customers that are interested in what they want to sell.

One problem is reverse mapping -- making it hard to scrape sites, etc. -- to get from retailer-specific identifiers to products. What if you could figure out a “crowd sourcing” way to build up the reverse mapping across a large set of retailers. With some distributed approaches, you could have competition among “evaluators” that map urls/ids to products. (aside) just modeling a “product” and what makes it unique and/or related to some other product is also a hard problem. Large online retailers devote a *lot* of human and other resources to each aspect of these problems -- product modeling, price comparisons.

User adoption is hard because of chicken-and-egg problems. You need a lot of data to attract users, which enables you to get a lot of data.

There are existing businesses that are operating in the interest of sellers to reverse engineer urls, etc in order to target customers - Behavioral Targeted Advertising. Sid suggests that perhaps these could be “rehabilitated” -- perhaps doing the same or similar things, but for the benefit of the customers who choose to participate.

Axiom aggregates data from public records, build a profile on you, and sell that data to others. If Axiom makes a claim about your data, e. g. salary, that might be a stronger claim that the individual making a claim about their salary. This might also be “rehabilitated” to be valuable to the consumer who *wants* to be able to have someone.