Technology

Lars Rasmussen: the brains behind Facebook's future

Mark Zuckerberg and Lars Rasmussen
Image caption Lars Rasmussen, right, looks on as Facebook boss Mark Zuckerberg discusses Graph Search

For a man that made his career out of helping millions of people find their way around, Lars Rasmussen is frank about his own navigational shortcomings.

"I have no sense of direction whatsoever!" he laughs, recalling the time he began work on his most famous project, Google Maps.

"I get lost in my own house - and it's not even a big house."

But while some would assume his poor grasp of getting around makes him the least ideal candidate to create the world's most popular mapping service - the Danish developer would strongly disagree.

"In mapping, where I have no sense of direction, my needs are likely to coincide with the needs of lots of people," he tells the BBC.

Mr Rasmussen left the search giant two years ago to join Facebook, a company he says "feels like working for Google when I first joined Google".

It is youthful and nimble, he says, with a more up-and-coming attitude than the search firm.

Given his experience with Google Maps, his current project, Facebook's Graph Search, appears a perfect fit.

"If you look at my desk... it's profoundly disorganised - which is why I can work on search," he jokes.

One engineer per million

The challenge is enormous for the ever-expanding Facebook - not just in users, but in personnel.

"Mark [Zuckerberg] has maintained this really interesting formula that we hire a new engineer every time we get another million active users," he says.

"I don't think any other company has that kind of ratio. It's crazy when you think about it."

Graph Search is perhaps Facebook's most ambitious challenge yet. Unveiled for the first time in January, it is attempt to get a grip on all the data it holds on you.

Image caption Lars Rasmussen is in charge of making sense of Facebook's huge data pool

It means users can use "smart" search strings, such as "photos of me and my brother in Ireland in 2006", accessing data that until now had been buried under the sheer mass of information uploaded.

If, as some analysts say, data is the "new oil" - this makes Mr Rasmussen something of a Saudi prince, thanks to his ability to make it more accessible, useful and profitable.

Soon, he promises, the search facility will become even more sophisticated.

"The big chunk of data ahead of us is posts - all the stuff that you see on the Newsfeed," he says.

"It's by far our biggest data source. The engineering challenge of building a search index that can manage that volume of data is big. We're well underway to making the system scale that far - we just aren't there yet."

It means a friend searching for "Friends who like football" could identify you if you had once written something like "Come on United!" in your feed.

What it remains unable to do - at least for the time being - is know how you feel about a subject. Not everyone mentioning Manchester United will be doing it with pleasant intentions.

"There are really interesting techniques around sentiment analysis," Mr Rasmussen says, "where people try to take text and figure out whether it's someone saying they like something or don't like something.

"But we haven't built that technology yet."

Privacy worry

If you're reading this and wondering why you haven't got Graph Search yet, it is with good reason.

Mindful of the inevitable - but valid - privacy concerns new Facebook tools spark, Graph Search is initially being rolled out to a small selection of users - many of them journalists.

Is this is sign that Facebook is worried about how users may react to Graph Search?

"Yes, very much," Mr Rasmussen acknowledges

The message is clear: when you get Graph Search, you need to check and double-check what information of yours is public. Pictures that previously required hours of scrolling to find, can now be drawn up immediately.

"We're rolling Graph Search out slowly - fast enough for the debate to take place, and but slow enough so people have a chance to look through their stuff before this comes out to a large audience," he says.

Days after the tool's launch, a Tumblr blog emerged highlighting some of the more potentially damaging searches that Graph Search had made possible.

Examples included: "Married people who like prostitutes", "Single women who live nearby and who are interested in men and like getting drunk" and "Islamic men interested in men who live in Tehran".

Such stories are embarrassing, Mr Rasmussen says, but a necessary evil.

"Without those occasionally annoying stories, people wouldn't go and check if they have appropriate data privacy settings.

"One of the things that made the company so successful was that we were one of the first that made people comfortable enough to share their real identities, their real lives, on the web. Before then everything was synonyms, avatars and so on.

"It's all about creating an environment where it's safe for people to share private things with the people they care about."

Wave of disappointment

On Mr Rasmussen's CV, between the phenomenal success of Google Maps and the promise of Graph Search, lies a monumental, and painful, flop.

Google Wave, which was three years in the making, was designed to be a tool to help teams collaborate - incorporating various strands of project management into one place.

After an initial surge of enthusiasm the service was killed off for good in April last year.

"It was a very painful failure for me," says Mr Rasmussen. "I've spent a lot of time thinking about it.

"I thought it had a lot of promise - and then couldn't quite make it work."

Unlike Maps or Graph Search, Mr Rasmussen felt that when it came to making tools to organise work, he was the best man for the job.

"When you look at my life, I spent a lot of time communicating, and collaborating, and co-ordinating. I run large engineering teams - I'm pretty good at this stuff.

"So it's ironic that my attempt at making a tool to make communication better, easier, faster… failed."

The lesson?

"In the future I should work on things in areas where I suck."

More on this story