How to go beyond technical features to choose the right product? Mix faceted navigation with folksonomies

January 26, 2009

When it comes to choosing the right product for your needs (let’s say a digital camera) among all the models available on the market, CSEs or retailers always use some sort of faceted navigation to let you drill down on their product database. And very often the facets proposed by those sites are limited to technical features (i.e: how many megapixels? 5x, 8x, 10x optical zoom?…) and price / stores that sell this product. But some companies try to innovate by enhancing faceted navigation with data coming from folksonomies, like WisdomTap or Buzzillions. 


As CrunchBase explains:

WisdomTap crawls the best social media sources, such as discussion forums, review sites, and blogs, to generate impartial product reviews, ratings and recommendations. It uses patent-pending technology to identify expert contributors in these forums, and extract their deep insights on products. Consumers are provided content that cuts through the clutter of a multitude of online product reviews and recommendations and helps them research products that are best suited for their specific needs.

Through crawling and analysis of product reviews, WisdomTap is able to propose a list of “intended uses” for each family of products (currently Digital Cameras and Mobile Phones only). You can then combine classic options (choose by brand, choose by price) with filters such as “Beginner“, “Professionnal“, “Outdoor“, “Bright Light“, “Wildlife“… 

WisdomTap - Camera search

WisdomTap - Camera search

So far, even if it looks promising, WisdomTap lacks coverage (how would such a system work on fridges?) and a proper product database (I need tech specs at some point to refine my search).


Buzzillions is also a review aggregator, much older and more established than WisdomTap. In 2007, , it started adding a similar concept as “intended uses”. As Peter Morville explains on his blog:

Pete Bell of Endeca turned me onto the fascinating, faceted, folksonomic innovations going on at Buzzillions.


Here’s what Pete had to say:

One of our customers just launched a beta of the most interesting integration of folksonomies and faceted navigation that I’ve seen so far. It balances the rigidity of facets built from a controlled vocabulary with the potential anarchy of raw folksonomies.Users can iteratively refine their search using any combination of controlled vocabulary terms and user contributed tags.

In 2007, those “intended uses” and the faceted navigation on technical features were displayed in different zones; tagclouds were used to illustrate folksonomies . In 2009, the user interface is leaner, both “intended uses” and technical features being displayed together as filters on the left-hand side column ; above results, a simple box highlights the type of users (for digital cameras: “casual photographer“, “photo enthousiast“, “semi-pro photographer“, “pro photographer“) for the category of products.

Buzzillions lacks the search approach used by WisdomTap, which reduces the flexibility to search into folksonomies ; however, the integration with faceted navigation on technical features is really nicely designed.

Dishwasher category at Buzzillions (Jan'09)


Easy :) Mix the power of searching folksonomies from WisdomTap and the nice integration with “classic” faceted navigation from Buzzillions… Also, add the enhanced product selection offered by YouTellMe (deep product definition ; define the most important criteria for you to buy), hire a very good interaction designer, and you can get the ultimate product selection tool :)

Commented by Rohit Chauhan
on Jan 27, 2009 / 10am

We are in the process of increasing product category coverage on Over the coming months, our engine will cover most of the important product categories including electronics, sporting goods, pet supplies etc.

Our algorithms are designed to work well for all types of products except those that are "subjective" in nature i.e. where the product ratings are based on individual user tastes - e.g. movies, music, books. So we can provide product reviews, ratings and recommendations for a large set of product categories (including refrigerators)

Another point to note is that we do deep semantic analysis of all the content we aggregate from the best social media sources, and generate detailed ratings for all facets of a product based on this analysis. This gives users a holistic view of products based on detailed ratings of features, as well as performance ratings for intended situations of use. Moreover, crawling and analyzing millions of threads of conversations enables us to automatically generate thousands of unbiased product reviews for every product.

Commented by Floyd-out
on Jan 27, 2009 / 11am
Buzzillions is really impressive of simplicity and clearness.

But I still have difficulties with all these clones of crawled reviews websites. The thing is that they consider each review at the same level of importance, whereas it's a power user review, a press review, or a simple user review.

I prefer the approach of alatest which is to mix user and press reviews but let the ability to separate the different scores. Even if their algorithm needs some tuning...

This is the approach I have chosen for my cinema website ( and I can't find today a best movie selection tool. objectively ;-)

Commented by nicolasleroy
on Jan 27, 2009 / 4pm
@Floyd-out - I tend to agree with you, even if you can have lame pro reviews and really great personal reviews from passionate people... It may work really great for movie reviews, do you think it would work great on product reviews?

Commented by nicolasleroy
on Jan 27, 2009 / 7pm
@Rohit Chauhan - Hi Rohit, thanks for those details regarding your algorithms. I'd really want to see how your engine can work on soft categories like sporting goods ; it also lets me think how we could mix faceted navigation and such "intended uses" on soft categories (like we have on Kelkoo: ) - the challenge is higher compared to electronics good as it's quite difficult to build product database / product clusters on soft categories and get reviews attached to them... Interesting :)

Commented by Vijay R
on Jan 28, 2009 / 6am
@Floyd-out - one of the unique things we're doing is to algorithmically gauge the expertise of the reviewer - so every review is, in fact, weighted by the expertise of the reviewer. The real world analogy is that you tend to trust the word of someone you "know" to be an expert...