Filter bubbles and the limitations of automation in selective search – 1
When we go online we virtually always face choices, and as the internet and its offerings have become more pervasive and inclusive, the range of options we have to chose from become daunting. I recently posted a blog entry: The Paradox of Choice, and Optimizing the Customer Decision Making Process in which I touched on the issues of how choice with insufficient distinction as to value and merit of choices becomes a barrier. Even when distinctions in type, value and merit are clear, on an item by item basis when the number of options that have to be waded through exceeds a comfort zone limitation, searching for the right thing becomes difficult – and with further increased scale impossible. I have written about issues of scale many times in this blog, and most recently from a problems and opportunities perspective with When Information Becomes Cheap and Ubiquitous Attention Becomes Rare and Costly.
The issues at play here, and attempts to offer choice without overwhelming, or burying the searcher in the unrelated have been ongoing as long as there has been a World Wide Web. And the search for selective availability that can in fact support being able to find anything out there online, is why search engines per se have become such essential online tools. And then Google, and a rapidly increasing range of other sites changed the basic underlying principles and assumptions of search with personalized search.
Netflix makes movie recommendations based on your history with them and what you have already seen, with their basic profile of your preferences fine tuned by your feedback response on specific movies viewed. But you can still easily search through and select from any of their complete list of offerings, including movies, television series and other items very different form anything you have viewed through them before. You can even change your mind and reorder something that you panned as terrible upon a prior viewing.
Amazon very actively and proactively offers purchase recommendations too, but once again, you can easily bypass that to search the site and all of its offerings as if a newcomer to Amazon and to online purchasing.
Google, on the other hand, provides personalized search without any alternative options and that is where the problem arises. When two different people make what is ostensibly the exact same search on Google, they now see perhaps completely different results, and not just in the first screen’s worth of search query results. Basically, the only internet visible to them through Google is whatever part of what is out there, that a Google algorithm analyzes as fitting your profile. And as far as Google is concerned, since December 4, 2009, search increasingly means being trapped in a mirrored inner-surface bubble.
There is a lot more to this story with this posting simply broaching a complex topic area. I will be coming back to it. Meanwhile, I offer a recommendation to a book on this set of issues.
• Pariser, E. (2011) The Filter Bubble. The Penguin Press.
I will add as a final thought, that while viewing customization and personalization can and does offer real value, it can only offer real value when it is optional and can be bypassed for a more open and unbiased view. Here, that would mean allowing an open Google search according to its completely un-personalized organic search priority algorithms, and with any pay per click or other different-model alternatives clearly identified as such.