Finally, some big news about Twitter which doesn’t involve the company being bought by Google Microsoft Yahoo Apple whoever this week. Instead, at a Cnet panel last week Santosh Jayaram, Vice President, Business Operations, mentioned that Twitter will begin crawling the links in tweets, and then indexing those pages.
As Techcrunch rightly points out, this isn’t a matter of Twitter trying to beat Google at their main strength, traditional search, it is instead a brand new paradigm, with the promise of human-influenced, real-time search results.
Little is known about the details but there were comments made about weighting links based on influence (a la Google’s Page Rank system), but no word on how that would be done yet.
One thing is certain though, this finally offers Twitter the chance to prove its real worth. All those big companies sniffing around for the past few weeks weren’t doing so as a result of the buzz around Twitter (despite Oprah‘s growing disinterest after an initial flurry of tweets, the number of new users continues to climb), but for the oodles of data Twitter has unprecedented access to. As with Google’s purchase of Urchin a few years ago, seeing how real people behave online is inherently valuable.
How Will It Work?
Short answer: we’ll have to wait and see. Longer answer: I’m not sure, but I’d like to see something like this. A real-time search engine,which would compliment Google rather than replace it, where real people invisibly influence the search results.
For example, a current Twitter search for ‘obama‘ has thousands upon thousands of results with dozens more each few seconds. Great if you want to know what people are saying, not so useful if you want to discover what exactly they are reacting to, as you have to wade through page after page to see all the links.
Or ‘manny ramirez‘ – you’ll see lots of comments about his drug violation, and many of them also have links. The trouble is that with URL shortening, you can see what looks like 10 different links all going to the same article.
Wouldn’t it be great to have this page split into two columns, one with comments and another with the most popular linked-to articles/websites?
Of course, it will also become subject to the scourge of the internet – spammers – but that is where the relative influence of the writer comes in.
It’s easy to forget now, but before Google search engines were filled with spam. Yes, it still appears throughout the search results now,even on Google, but nothing like it was back in 1997/8, and that is largely because of Google’s PageRank system. For a detailed mathematical explanation see here, but basically Google used links to a webpage as a way of measuring its popularity,but with the very important caveat that not all links are equal.
If I run a golf course for example, a link from PGA.com would be far more valuable than a link from your old Geocities page – PGA.com is a popular site, and in a related field. A similar thing could be done for Twitter search, a link from ESPN columnist and professional RedSox fan Bill Simmons to a Manny Ramirez article would be worth more than one from Oprah, even though she has more followers.
And of course, Twitter search wouldn’t have to be limited to websites either. As with Google’s move towards universal search, it could easily incorporate videos, mp3s, photos, and anything else that people are discussing or linking to.
Something like this could be hugely useful, if implemented correctly. Despite Google’s success, the missing piece in their search has always been a human element. In fact they incorporate the Open Directory Project listings into their search algorithm as a way of helping to inject some personal judgment.
A search engine that shows exactly what is happening around the world right now, as discussed by real people? I can’t wait!
So what do you think? Am I missing the point? Is this another potential ‘Google-killer’ that will die on the vine? Please share your thoughts or comments.