Startups Take on Amazon Reviews


While low prices and huge inventory are the main attraction at, the site’s reviews have emerged as an essential online resource. Now, a new generation of start-ups wants to take on the shopping giant by spreading reviews across the web.

As PowerReviews CEO Andy Chen puts it, he’s building a “next-generation Epinions.” His competitors in the distributed reviews space include Bazaarvoice and European-oriented Reevoo. Today, a new start-up, Ratepoint, is throwing its hat in the ring.

Boston-based RatePoint has raised a little over $1 million from Prism VentureWorks, .406 Ventures, and its founders, who were part of the team at GeoTrust, which they sold to VeriSign in September.

The start-up has gone from concept to launch in four months, an impressive feat — but then again, that might just show how wide open this category is. RatePoint makes a toolbar (as of Sunday night, only available for Internet Explorer) that aggregates user ratings. Toolbar users will see reviews weighted to emphasize other users with similar tastes.

“A five star for me is not always a five star for you, it might be a three star,” explains founder Chris Bailey. For now, the user reviews are only of URLs, so the tool the service most closely resembles is probably StumbleUpon. Product reviews are in the works.

PowerReviews, which has $6.25 million in funding from Menlo Ventures, Draper Richards, and company founders, is taking a slightly different approach, though its earthy green and cutesy star design is pretty similar to RatePoint’s. We’re pretty sure PowerReviews came first.

The company offers review software and management for retailer web sites — for instance, Walgreens — and distributes the reviews across its 80 customer sites, taking a cut of ad and product sales. Its staff of 18 does a diligent job of quality control.

Until now, PowerReviews’ reviews have only been available on customer sites, but the company plans to launch a portal in the next few months. Since the review templates are extremely specific, the portal will provide some interesting side-by-side and tag-based comparisons. It will also expand a revenue stream of sponsored listings (though sponsorship does not affect search ranking, says Chen).

I saw an early version and liked it, though I felt like the site could really benefit from being mashed up with price comparison, local availability, and color tools. How many shopping web sites are you going to go to that don’t actually sell products?

Both companies’ ideas are interesting, but my hesitance towards installing yet another toolbar makes me prefer PowerReviews’ approach. Reevoo uses primarily email surveys (to ensure customers have actually bought a product), which also seems a little too onerous.

By staying focused on reviews, the companies are ensuring that they always stay relevant to consumers. However, they’ll only become useful — and profitable — when they elicit a whole lot of participation. Amazon may be in need of some competition on the reviews front, but it’ll be hard for any site to emerge from the pack.


Kingsley Idehen


Unless I am missing something, I note that this service, like many Web 2.0 services, is loathe to share data. Of course many assume that Open Data Access is inversely related to business model success, but this remains a fallacy based on a lack of understanding of what Open Data Access delivers when you engage the Semantic Web vision. is an example of an Open Data Access model based Review Service. I can query this data from outside the network while the network continues to grow and provide value added services.


To bad their isn’t a way to aggregate all reviews into numuerous toolbars as an extension that would fit all toolbars. I think something like rankings would be cool to see next to search results from multiple bars like toolbar does.


As Bailey says, one person’s five-star is another’s one. The five-star rating is ubiquitous and so a nice standard, but almost to the point of losing any meaning. I’m always amazed at sites that put two products “stars” on equal footing only to write in small print that the one-star product is the result of one review where the reviewer is complaining that the product lacks a feature it never claimed to have in the first place. Maybe these models should require a minimum number of reviews before posting a star rating. Anyway, these are all things that throw wrenches into the works of the aggregators. Our solution? Skip the algorithm and let human editors read everything they can, make informed judgments about information quality and draw a short list of the best products.

Ben Watson (Thomson NETg)

What I really want is for to open up it’s rating/ranking/wiki/reviews system so that it can sit on top of anything, not just products. Amazon wins by reaching into a larger audience of potential users and I win by being able to leverage their investment and place (in my case) the value of my online courses in context.

Like Om says, I need something that is easy for users to use and that already has a large audience that can leveraged/cross-pollinated. Even better is if Amazon leveraged the S3 to independently store ‘my’ reviews while they are able to present them in the aggregate.


Ben Watson – VP, Collaboration
Thomson NETg

Vaibhav Domkundwar -

Hi Liz:

Interesting post. Reviews and ratings continue to be central to decision making for most users and there is definitely an opportunity to deliver a “dead simple” experience to the users, as Om commented.

We have taken a different approach at iNods (, where we are aggregating, consolidating and presenting reviews and ratings in a structured format to the end users. While Amazon has the largest amount of valuable review content, we believe that it will be difficult to replicate that model on a single site or on a syndicated platform. Also, most of the “new” user reviews and feedback are appearing on personal sites and blogs and so on. iNods aggregates and structures all this content, including Amazon’s, and users just have to search or browse to a product thet are interested in. Let me know what you think.


lawrence coburn

Nice summary. I think PowerReviews has a neat model, and it seems to me that they’ve already done the hard part which is to integrate their free ratings engine with a ton of retailers. Reviews aggregate at a higher rate when they are point of sale, as opposed to on a third party site.

Also, don’t forget about RateItAll in this space. RateItAll is a long tail consumer ratings site in which consumers create, promote, and manage their own top ten style lists of products, people, places, etc, which can then be posted on their blogs or social networking pages via a variety of widgets. Publishers share in the revenue associated with their contributions, and reviewers own all of their reviews, profiles, and rating lists – which they can remove at any time.

We are unfunded and running on cash flow.

Om Malik

I think the trick is coming up with a way to make reviews dead simple for end users – in fact turning the review process into a game or something – that is how you can get people to participate more. Sort of like multiple choice questions in a SAT exam, except with answers people might actually be thinking about.

Eric C.

This is the chicken or the egg problem no? Rate of reviewers is pretty low compared onlookers. So, it is hard to get the data without already having a lot of traffic. That is why sites like amazon/netflix can be successful.

I find the challenge to be giving relevant and useful information to the onlookers before you hit the saturation point needed to get good amount of user contributed data. Its a challenge I am struggling with on my project as well.

Kris Tuttle

Reviews seem more open than one would think. When they exist they are valuable as Amazon knows. However I recently discovered that all reviews are sole property of Amazon which makes many prefer to post them on a blog or elsewhere to keep control and copyright.

Nobody wants another toolbar but having an aggregation of quality review content would be great if it could be done. Getting the content however seems to demand some creative methods and incentives.

Comments are closed.