Netflix, Qwikster, and the Potential Failure of Search UX
Yesterday, Netflix announced the separation of its DVD-by-mail division into a new entity called Qwikster. The new company will have a separate website and focus entirely on mail delivery of DVD and video game rentals. It also will, in a few weeks, have a new website.
In splitting the previously Siamese twins, Netflix also decided to make its users create new user accounts and undergo separate billing for its two companies. From a business perspective, this makes sense, as it will allow each company to focus on its specialties and to have a separate P&L statement.
However, as pointed out in the comments to CEO Reed Hastings’s blog post (search for the comment right before Hastings’s reply of “ouch”), the team didn’t think of important potential scenarios in the new dual entity system. Netflix is a Solr user, but here is a use case that points out that mere implementation of Solr does not a great search experience make. The poster queries about a simple use case: what if a user of one system searches for a movie and it’s not available on that system. Will the search interface allow the user to see if the movie is available on the other system? Thus, let’s say that I search for The Amazing Race on Netflix. It’s not available by streaming, but it is available by DVD. However, as Hastings sheepishly points out, in this case, “You’d have to search the second place if we didn’t have it in the first place.”
Oops. Cue user discontent.
This is not the first time that enterprise search designers have failed to take into account pretty key use cases of search when rolling out a search engine. I see intelligence analysts having to struggle with poor search user experience all the time, unfortunately.
How should enterprise architects work to mitigate this scenario? Conducting personae and scenario analyses. Identify the typical personae who will be using the website and the potential scenarios in which they will be using search. Search is more than just a white box with some text that you enter, click on search, and magically get results. By thinking through the use cases, then marketers will have a better set of requirements to pass along to the IT team responsible for implementing search.
The IT staff of Netflix is technically correct. They have a search engine that works. What they might not have, as Hastings had to publicly admit, is a search engine that works well.
(Note: Opinions expressed in this article and its replies are the opinions of their respective authors and not those of DZone, Inc.)