March 04, 2009

Likaholix Launches Recommendation Engine Based On Social "Liking"

New services for people to connect to peers and share their interests are cropping up around every Web corner. Some, like Facebook and FriendFeed, have found strong, growing communities, where people find things they care about, and discuss those items. And as FriendFeed has discovered, the simple act of "liking" an item and enabling comments, helped the site differentiate itself from practically every other aggregator out there. This morning, Likaholix, sprung from the minds of some former Googlers, follows suit with a site dedicated to just your likes - whatever they may be, but adds on suggested items, recommendations, and even experts.

(Skip this post and get one of 200 invites: here)

No matter how many new lines of business Google finds itself in these days, its core value has always been search and retrieval of information. That mantra has been seen in sites developed by its former employees, like FriendFeed, with its integrated search, and now, also with Likaholix, who takes things a step further, by not just crawling a database, but auto-competing search terms, as does Google Suggest. And these auto-suggestions come packed with a ton of books, restaurants, and even movies and music - in what may be a vain hope that you won't see the same few dozen self-proclaimed social media experts lauding the latest technology update over and over.

Likaholix's Search Engine Auto-Suggest Feature In Action

Likaholix's Search Engine Results

Likaholix even suggests items that nobody likes, guessing that somebody will, and eventually, they would be selected.

Surely Somebody Finds South Park Funny?

As Likaholix is all about what you like, you're credited for being the first to like any item. On your profile, not only does it show what things you like, and who you like, but it shows how many times you were the first to like an item, and introduce it to the community.

Searching for Twitter Shows It's Not Yet Been Added

And should you be the first to like something, Likaholix helps you fill out your data profile with a description, and even images or videos pulled from Google to help round out your item.

Introducing a New Like to Likaholix

A Friend's Profile on Likaholix

Like with FriendFeed, your stream of likes and comments are found on your profile. You can also see how many different people liked an item, and who they are, and see recommended items based on what you've told the system you like. You can also view how many other users of the service you like, and who likes you.

Find yourself an expert in something? Claim you are with Likaholix - whether you're one of the previously mentioned social media experts, or if you have a love for science fiction, SEO or even art. Claiming expertise puts a star next to your profile, and will display your name with a star beside it anywhere it is mentioned, including on the list of contacts your friends have.

Tudor Bosman - Science Fiction Tastemaker on Likaholix

The core focus of Likaholix is to recommend to you more things you might like, based on things you say you enjoy, and those items your friends do - developing an automatic recommendation engine of sorts. For example, this morning, I was told I probably would like TiVo, because Sanjeev Singh, who I have liked, says he likes TiVo as well. In this case, Likaholix is right. It's not always right of course. I have less of an interest in (a fashion Web site) or Downtown Brown beer (considering I don't drink), but I assume that as I like more things, and my friends do as well, this list can get better honed.

And as you like items, you can even share them in the usual places, like Twitter, Facebook and FriendFeed - extending the reach of the site.

Likaholix is opening up today in a private alpha mode, as they say. Existing users, like me, can invite friends to try out the service. The user interface is spartan for now, but the team has no doubt been working on the recommendation engine for starters, and a sharper GUI can come later.

You can find me at: Be one of the first 200 to use this code, and you can get in as well: