Twitter’s search sucks. That’s why Algolia re-built it (better) from scratch.

Jan 10, 2014
Vote on Hacker News

download (3)

Twitter’s search sucks. There, I said it. My default action when I want to find a person or company on Twitter is to open a new tab in Google Chrome and search “Name” – if they’re relatively well known, I include ‘Twitter,’ though it’s not usually necessary, given the social networks high traffic and ranking in Google. I’m consistently disappointed with Twitter’s search, where searching for keywords on the desktop, or when I try to find someone via their iOS app; I can’t imagine I’m the only one.

Search as a Service startup Algolia seems to be fed up with Twitter’s sub-standard search as well; the startup, which raised $1.5 Million from Index, Alven & Point Nine last October, explained in a blog post how they used Twitter’s API to pull over 30 Million twitter accounts, their names, descriptions, and @user mentions, in order to rebuild Twitter’s database (or at least a portion of it).

Using typo-tolerance and ranking, two features of Algolia’s product, they were able to create a Twitter search that arguably works better than Twitter itself.

download (2)Search “ruedebaguette” (typo intended) on Twitter, and it’ll load results based on its assumption that you meant rudebaguette. search “ruedebaguette Twitter” on Google and it returns a beautiful list of twitter users who have incorrectly tweet mentioned @Ruedebaguette. Using Algolia’s Twitter Search, I had RudeBaguette as an instant result after typing Ruedeba (Algolia’s search produces results after every keystroke).

Of course, the stand-alone search is too inconvenient to use as replacement to Twitter’s built-in search; however, it goes without saying that, if Algolia can build search that produces smarter results – – their blog shows that search results for @adele or @bieber do not turn up the respective singers, but instead turns up fan accounts starting with the names – then Twitter certainly has no excuse.