(Page 2 of 2)
Yolink does not rerank search results, but it does rank the information it extracts from the page to determine what may be most useful to the user. On a Yolink results page, the user can click on extracted data to see a more detailed preview without navigating away from the results page.
TigerLogic originally built Yolink as a browser plug-in, but the company changed its strategy to provide an application programming interface (API) that lets publishers and website owners integrate the technology into their own sites. The company offers this service to website owners for a range of prices based on how many searches its users perform each month. The service is free for noncommercial sites with fewer than 2,000 searches monthly, and $1,500 a month for commercial sites with 2,000 to 25,000 searches a month.
There's a growing sense within the search industry that end users are frustrated with the clutter of results they have to wade through, says Greg Sterling, a senior analyst with Opus Research, a market research firm based in San Francisco, and a contributing editor for the website Search Engine Land. Sterling sees Yolink as an effort to reduce the time it takes to decide which link to click. "They have done a nice job of evolving a complement to search," he says.
The tool isn't competition for Google or Bing, Sterling says. Instead, it competes with similar search enhancement efforts. "The challenge for anyone in search is to get attention," Sterling says, "to create an experience that adds something, and that really goes beyond Google." He thinks publishers and website owners are much more likely to understand how Yolink could improve the search experience on their sites than end users. Sterling believes Yolink's API strategy will work much better than offering a browser plug-in, which users would only download if they already understood the likely benefits.
The Chrome browser will soon silently fetch pages as you scan search results so that they load without delay.
Search queries aren't always better than traditional trend-spotting methods.
A Yahoo Research tool mines news archives for meaning--illuminating past, present, and even future events.
While I applaud anyone working to make search more semantically sophisticated, there is never going to be a great solution to the problem of irrelevant search results for two reasons.
The first is that there is an entire industry of very intelligent people actively working to sabotage search result relevancies for the profit of their clients - the SEO industry. This is an arms race without an end.
The second is frankly, that most searchers are stupid, umm, maybe I should say naive. They have such limited command of their own language that they don't perceive how the word (often searches are a single word!) or words they type could have multiple meanings (English is notorious for this). There is just no way a search engine can be smart enough to read their minds. The only solution for this is generally considered unacceptable by many people, which is for the search engine to track their individual search/click behavior over time to personalize its interpretation of their searches in order to gain a contextual understanding of what they want (e.g. whenever I search "dolphins", I'm always interested in the animal, never the football team).
As a librarian who struggles to improve our little search engines (eg library catalog), I am very grateful that I don't have to deal with someone actively trying to manipulate my indexing data like the SEO folks. And as a reference librarian, I see university level students type in the most unreasonable search strings all the time, unreasonable that is to expect that they'll get relevant results given the intent that they just articulated to me as we work together. Library science folks have pretty much reached the conclusion that we have to offer them what we call a faceted approach, which others may think of as a guided drill down, to help them refine their first unhelpful set of keyword search results with suggestions of new terms and other limiters (eg date of publication, language, geographic region) to narrow in on what they want. You see this on Amazon all the time, those boxes on the left that help you narrow your product results by brand, price range, range of GB size (for storage devices), etc.
I'm surprised Google hasn't found a way to do this yet, although since they (unlike librarians and Amazon) are working with an entirely uncontrolled index set, it's an enormously more difficult problem for them, but still I expect they'll figure it out, hopefully soon.
Manufacturing in the United States is in trouble. That's bad news not just for the country's economy but for the future of innovation.
On Twitter
Become a Fan on Facebook
Subscribe to the Feed
smithsomian
182 Comments
not a competitor to Google?
don't quite buy that. Google is certainly working on similar enhancements to their search products, so whether these chaps see themselves as comptetition to Google or not, I would imagine Google will see it that way.
Reply