One of the problems with user reviews for publishers is that if you reach critical mass then they can become unwieldy to navigate and make sense of. Trip Advisor is my favorite “whipping boy” in this discussion. Scores of (often contradictory) reviews potentially make it difficult to get to the “bottom line.”
Few publishers are in that position. But in selected markets, such as San Francisco, Yelp is one of them. To address this the company has added a summary feature that shows the distribution of ratings and graphs them over time. The following is from a restaurant called “Sushi Bistro” in San Francisco, which has 180 user reviews:
If you click the “rating details”‘ icon, you get the following charts:
This is a very helpful presentation of the data, showing the distribution of reviews and how they’re trending over time.
You can also sort reviews by several criteria, which helps users navigate through the throng (180 reviews in this case).
Another feature recently upgraded is its collaborative filtering: people who viewed restaurant X, also viewed restaurant Y:
This is both helpful to the user and a way to drive more page views.
___
Marchex’s OpenList also has filtering and sorting and has an alternative way to summarize reviews:
October 5, 2007 at 6:23 pm |
My friends at Summize (http://www.summize.com) have a useful approach to these issues, although they haven’t yet applied it to local — despite my requests.
October 5, 2007 at 6:30 pm |
Oh — and WSJ.com had an interesting article about the statistical perils of summarizing star reviews today:
http://online.wsj.com/article/SB119153813516649500-email.html
(Will that link work? Hope so.)
October 5, 2007 at 7:16 pm |
Yes, it’s a nice feature. Over at YellowBot, while using our own product, we noticed the same issues and added this feature in months ago.