Collecta

» see collecta.com as it lived, on the wayback machine

~ the obit ~

Collecta opened to the public on 18 June 2009 from an office on 9th Street in Santa Monica. Gerry Campbell had run search and content technologies at Reuters, and before that he had been a senior vice president of search at AOL. His chief technical officer was Jack Moffitt, who held board seats at the XMPP Standards Foundation and at Xiph.org. The company was therefore built by one man who knew how to sell search and one who knew how to move messages around.

What it sold was a search engine that refused to rank. Results arrived in the order they had been published, newest at the top, and there was no relevance score anywhere in the product. Campbell described this as throwing aside the common conception of what search is all about. The specific thing he had thrown aside was the ranking function.

The index was assembled out of English-language blogs and Twitter, and that alone explains why nobody in Turkey ever had cause to open it. Collecta was built for people whose work involved watching a single word move minute by minute, and in 2009 that was a job description that existed in a few American newsrooms and almost nowhere else.

~ the rap sheet ~

~ push instead of pull ~

The engineering argument was about direction. Every other search engine went out and asked for content, crawling pages or polling feeds on a schedule, discovering things some interval after they had happened. Collecta had publishers push to it instead, over XMPP, the protocol that carried instant messages. WordPress pushed to Collecta, and so did a weather station run by Iowa State. Twitter came in through the API, because Twitter would not push.

Campbell put the difference plainly: with RSS or anything else you have to go out and ask for that information, and what Collecta had was push feeds. The effect was a results page that moved while you watched it, with no refresh and no interval, which in 2009 was an odd thing to sit in front of.

~ the year everyone wanted now ~

Collecta opened a few months ahead of the category's best moment. In October 2009 both Google and Microsoft signed deals to put Twitter into their results, and for roughly eighteen months real-time was what search was supposed to be turning into. Collecta had company in the category, with OneRiot and Scoopler selling versions of the same promise, and the money arrived on schedule. Dace Ventures and True Ventures put in 4.7 million dollars in the spring of 2010.

~ what the ceo said ~

On 19 January 2011 Collecta shut down the search business, including the site it ran on and the widgets it had handed out to publishers. The company itself was not closing; the search engine was. Campbell said there would be a pivot to related work he was not ready to describe, and Jack Moffitt had already gone. A reporter asked him whether building a real-time search engine was a viable startup business. He said no.

What replaced the product was a page with an email box on it. By February 2011 the homepage had fallen from roughly 4,800 bytes to 1,755, and by that July it was down to 841. The last version carried an invitation to sign up and stay in the loop, printed under a line from Benjamin Franklin about not squandering time, since time is the stuff life is made of. An engine built to tell you what was happening this second spent its afterlife quoting an eighteenth-century printer on the subject of wasted hours.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »