recommendation engines

Nara raises $4m to build a better recommendation engine

When it comes to finding the perfect restaurant, hotel or entertainment venue, chances are you turn to one or more user reviews sites. After all, if you’re going to trust anything, why not trust testimonials from a business’ past customers?

User reviews, of course, aren’t perfect, and sorting through them can require a lot of effort. Hence the effort many companies are making to build recommendation engines that use computing power to tell you where you should eat your next meal or go to have a good time.

Start me up! A profile of nToklo

Ntoklo logo Reports regularly show that nothing drives online retail conversions like recommendations from friend, but some businesses have struggled to implement a simple recommendation system.  

We spoke to nToklo, a new product that aims to solve this problem by making the recommendation model truly social.

Hunch’s hunch: less traffic will produce a more useful service

For most consumer internet startups, more is better when it comes to
traffic. But Hunch, the recommendation engine co-founded by Flickr
co-founder Caterina Fake, has a hunch: when it comes to traffic, less
is actually more.

So yesterday Hunch made a drastic change to its service, which comScore
estimates receives approximately 750,000 unique visitors each month: it
cut off access to users who aren’t registered and logged in. Fake told
TechCrunch that she thought “traffic will plummet” in the wake of this,
but that “users who are using the product will have a significant lift
in the quality of results.