Startup idea: Product attribute database

There are over a million online retailers in the US alone and over 2.5 Million worldwide. Many are in categories that are large and well defined (apparel, electronics, books, etc). If you are a online buyer one of the many things you want to do is to research a product well – understand the features, options and compare it to other similar products.

These are defined by what I call product attributes.

Comparisons & reviews are largely subjective and prone to long tirades and endless sentences without getting to the point. Here’s an example. Notice that current attributes that are already stored by Yelp include time the restaurant is open, expected attire, etc.

Those are some of the things I’d like to know.

A large number of things I’d like to know are not really comparable.

They are mentioned in the 88 reviews provided by end users. Taste of the food, visual appeal of the food, softness of the bread.

Reviews that are unstructured are a pain – to sort, filter, compare and review.

There are many who claim that the product attributes for products such as cameras and mobile phones are fairly complete and those are problems already solved.

I think that’s broken thinking.

If you look at how people search for cameras, many (not all) “lay people” dont search for HD pixel density, dual core Snapdragon processors etc.

They search for “how to take good pictures in the dark with your <favorite phone>” or “how to record a live band in <favorite phone> without the background noise”.

The attributes that customers want are those they use the product for. Unlike the specifications and features that manufacturers (or producers / service providers) build them with.

I think if you build a product to classify the attributes that matter for every product (start small and build by category) with a combination of technology and crowd sourcing (or any other mechanism), you will build a valuable company. 

Not to mention that its highly likely that Google, Yelp, Microsoft or others will buy you.

Advertisements

5 thoughts on “Startup idea: Product attribute database”

  1. This is quite a coincidence. I have been working on this – how to crowdsource an attribute and let users associate any attribute with the object they wish. There are a couple of challenges regarding storage and avoiding duplication of attributes. Will ping you once I have the prototype working.

  2. findthebest.com the best compares features and have reviews being added to various sources like CNET. Not exactly the same thing but it comes pretty close.

  3. Hi Mukund,

    This is indeed a very good idea. The next step the Sentiment technology has to give is precisely this: identify not just the overall sentiment of an opinion, review, etc, on a product, but also which particular “facet” of the product is the opinion about (“I love the cases of the iPhone but the price is still very high”). We read some time ago an article discussing Amazon reviews with this concern in mind (http://moz.com/blog/how-i-wish-amazon-reviews-worked), and we also talked a bit about it (http://www.bitext.com/2013/04/amazon-reviews-i-want-semantics-i-already-stars.html). So I agree it’s the way to go.

  4. This is indeed a very good idea. The next step the Sentiment technology has to give is precisely this: identify not just the overall sentiment of an opinion, review, etc, on a product, but also which particular “facet” of the product is the opinion about (“I love the cases of the iPhone but the price is still very high”). We read some time ago an article discussing Amazon reviews with this concern in mind (http://moz.com/blog/how-i-wish-amazon-reviews-worked), and we also talked a bit about it (http://www.bitext.com/2013/04/amazon-reviews-i-want-semantics-i-already-stars.html). So I agree it’s the way to go.

Comments are closed.