What could real-time microblogging search do for Google?
From a user experience point of view, the current search does not really feel real-time mainly because the results don't update as they do on Twitter Search. Real-time search would also make the Google search engine more aware of the context of the search query and hence be able to deliver context-aware advertisements.
Yet micro-blogging introduces an interesting problem that Google would need to solve first. Conversations on twitter are threaded and posts can come down to a short
user2: @user1 yes its amazing
which means context is spread across all the conversation's threads. There are several applications online that give you access to a threaded view of Twitter - something that Twitter does not have but it should. Any real-time microblogging search needs to support threads to be able to deliver more relevant content. By nature search queries are really questions which works well because conversations usually start with questions or an opinion. When I search 'mac mini' on a real-time search I'm not very interested in Apple's Mac Mini product page, I am more interested in knowing what others are saying about it right now. So If my hypothetical Twitter post above was in response to:
user1: do you like your new mac mini?
searching for 'mac mini' should ideally be able to show that threaded conversation even though the words 'mac mini' are not present in user2's response.
Real time search and Search Engine Optimization
Not so long ago - when people read physical news papers instead of online news - if someone wanted to know what has happened in the past 24 hours they would buy a newspaper. The news paper is similar to real-time search, it will inform you of new events around you and in the world. Also not so long ago - when people went to the library to get access to newspaper archives instead of online archives - if someone wanted to learn about an event that happened 10 years ago they would visit the local library and pull out archived newspapers that mention that event. Although current search engines can index pages very quickly, it still feels like archived news - in fact search engine result pages are really archived news.
In a real time search engine pages need to be ranked a little differently that how they get ranked right now and hence SEO practices need to change to accommodate these changes. Because conversations on a microblog like Twitter are threaded user2's response 'yes its amazing' would not rank at all on a search for 'mac mini' although that response is probably what the user wanted to see i.e. mac mini reviews. Maybe another user responded:
user3: @user1 its great, i installed tweetie, evernote and am in love with front row
That too should show up on a real time search since it is showing different applications one can get on a Mac. Threaded conversations on microblogs should be looked at in the same manner comments are viewed at the bottom of a blog post - except these comments are not nicely grouped together with the original post on Twitter.
This introduces new problems, since conversations can have nested threads:
user4: @user3 yup tweetie is the best Twitter application and front row is just too sexy
and then loops:
user4: @user1 I use my mini as a media server at home
Real time search is Social Search
Information is being added to the Internet today at a much faster rate than before the Social Networks age. A lot of this content is being added by individuals all over the world on dozens of social networks like Facebook, MySpace, Twitter, Blogger, Youtube, etc. In my opinion this content is the most useful to be presented in a real time search engine. Real time news can still be delivered through the participation of news organizations in social networks - as they do today.
Real time search would also allow the search engines to deliver truly personalized content and targeted advertisements. Google recently allowed users to create Google profiles and link it to your online identities across the many social networks. Search today is not aware of "where you are right now" so how can it truly direct you to where you want to go? Consider the Google search page as a map, and for Google to be able to deliver relevant answers to you, it needs to know your "location". I don't mean location as in a physical location with a latitude and longitude, I in fact mean context but for the sake of the argument I am using "location". So imagine the Google search page is a map and you are looking for "mac mini". Google goes out and gets you the most relevant results according to Google for that search term and plots it on the "map" i.e. the search engine results page. If your search term is broad, then your "map" will just contain "locations" all over the place - i.e. Google is unaware of your search context and depends on you to point it to the right context. With real time search and Google profiles, Google can make an educated guess about that context i.e. Google is able to tell what your current "location" is. Real time search on Google can now become a "GPS" and route you to the most relevant "locations" to you. By being able to search your threaded conversations on Twitter for example about the "mac mini" Google can customize the real time results page to contain content that you have discussed with others. Now my search for "mac mini" would produce results on "tweetie", "evernote", "front row", "media server" and "mac mini" - all keywords in the threaded conversation above.
Now even the advertisements on that page can be personalized instead of the ones I see now including "mac mini store" and "mac sale". These ads are not relevant to me since I already have one - so there is absolutely no compelling reason for me to click these ads. But on the other hand if the advertisement showed "evernote sale 50% off pro account today only", first I would be blown away that AdSense showed that to me, and second I am more likely to click it because I have the free version of Evernote.
No comments:
Post a Comment