LinkedIn beefs up search, but avoids Facebook methods
The social network discusses its new search technology that takes a page from rivals, but for a different purpose.
LinkedIn is searching for a better way to connect its customers.
About a year ago, engineers at the social-networking company for professionals realized the technologies powering search for different parts of its website just weren't going to cut it.
The problem: LinkedIn's user base was getting so large -- now more than 300 million registered members -- that the group of different technologies built to allow users to search for jobs, groups, and other things was starting to get stale and in need of change.
"We really hit a wall," said Asif Makhani, head of search engineering at LinkedIn.
Instead of relying on a different piece of search software to power each individual function, the search team, composed of 15 employees working on infrastructure, decided to create a new system that can work for all of LinkedIn.
So, the company created a new search technology it calls Galene, after a Greek goddess personifying calm seas. Indeed, that's exactly what it's supposed to do: make it easier for LinkedIn's engineers.
LinkedIn users will get access to the new technology over the next few months, though the changes will mostly happen under the hood. Search will be twice as fast, LinkedIn says. And there will be one new feature: instant results and suggestions, offering hints while users type. Type something like "Jav" into the search bar, and LinkedIn will ask if you're searching for a job writing code in the Java programming language.
LinkedIn isn't the first company to revamp its search product. Google has revamped its search product several times, including live-type search -- called "Google Instant" -- introduced four years ago. Apple was even granted a patent for technology that searches across the Web, on the device and in apps at the same time.
But new search technology isn't an automatic crowd pleaser. Last year, Facebook unveiled "Graph Search," a more advanced way to find friends, products, and events across the service by typing queries such as "College classmates who like Coldplay and X-Men movies and live in San Francisco."
But the promise was grander than the delivery. In practice, the product was complex. It didn't always work, and it turns out that pulling together all the information about a user from the company's many databases has proved difficult, the company said in a recent interview with BloombergBusinessweek. Graph Search also was developed for desktop browsers, and had to be "redesigned" for phones.
LinkedIn believes it is avoiding many of these struggles by focusing much of its work on the underlying technology at first. Merely making search work better will likely be an instant crowd pleaser, said Sriram Sankar, an engineer at LinkedIn who also worked on Facebook's search technology.
"If you go sit with a recruiter using LinkedIn, you see them doing complex searches you don't expect them to use elsewhere," he said. "You cannot imagine LinkedIn without search."
Updated at 12:15 pm PT: To clarify Asif Makhani's job description.
