Friday, October 5, 2007

My Computer Doesn’t Know Me Anymore

So what happens when personalized web promotions or on-site behavioral targeting tactics miss the mark? Technologies designed to improve the online experience by making it more relevant based on user interactions and predictive modeling can have negative repercussions leading to frustration and abandonment when assumptions are false.

I started thinking about these technological miscalculations when my Uncle explained that his Netflix account persistently recommended foreign films based on selections he made upon initiating his membership. He likened the recommendations to unwanted advances from a foreign temptress and has been anonymously patronizing his local Blockbuster ever since. Granted, Netflix offers several methods for customers to refine preferences and continues to hone in on a users’ specific taste through ratings and explicit information requests. However, we as humans are inherently lazy, and often times a laborious process of correction is more than we wish to invest, especially for the casual user. While Netflix is arguably leading the market in terms of allowing users to refine their interests and making effective collaborative recommendations, other methods of targeting are more difficult to correct, especially for non-savvy web users.

Take for example my Mom; years ago she diligently researched her doctoral thesis on minorities in education by venturing to numerous online and offline sources covering diversity, multi-culturalism and blacks. After some time she began to receive emails offering Black Education Journals and other periodicals at no charge. In this case the targeting was correct in identifying her interests and she accepted the offers, but the implicit assumptions of my Mom’s ethnicity were wrong. So, when subsequent emails started to arrive inviting her to join black singles networks, the targeting couldn’t have been further off base. The problem was that my Mom had no recourse for correcting the profile of her online self and still receives off-target emails and misdirected online promotions to this day.

In my own experience, I was issued a company laptop that previously belonged to another employee. Upon my first several Google searches, I was puzzled by the fact that all of my results were coming up from the UK. I was accessing the site through a North American IP address, yet after numerous searches Google was determined to prioritize UK results. I attempted to correct the situation using Google’s interface, but at that time found no course of action to inform the site that I was in the US and seeking US-based information. Try as I might, there was no correcting the problem and eventually I deleted my cookies (against my analytics-induced nature) and acquiesced to starting anew.

So, the moral of this series of observations is that sites providing personalized or targeted content must also provide a way out. Do this by offering methods for visitors to explicitly state preferences and by empowering visitors with clear messaging on how to turn off or modify behavioral targeting. In an age where tuning out unwanted advertising and content is increasingly prevalent, don’t you want to accurately target those that are listening?

No comments: