When I was speaking to IAC’s Dinesh Moorjani about the Citysearch Android app late last week, we discussed the idea that reviews were moving toward “commodity” status. While this clearly isn’t true across the board it’s already be true in the restaurants and hotels categories, for example. Take this restaurant review on Yelp (click to enlarge graphic):
There are 187 reviews of this (very good restaurant) in San Francisco. In other than “high consideration” scenarios, I’m only interested in the reviews summary (or maybe the two most recent reviews). Thus Yelp’s trends/summary tools become my focus rather than the reviews themselves.
We can discuss and debate when you might look more closely at reviews (dentist, roofer, etc.); however the paradox is: the more reviews there are the less they matter in a way.
It used to be very difficult to get reviews — and still is for many sites — this is what helped build Yelp and its brand: breadth and density of reviews. However when there are 100, 300 even 500 reviews people don’t have time to go through them and they (individually) have less value. Citysearch argues that editorial content and enabling the SMBs to “have a voice” on the site become differentiators in this new environment. (Yelp allows businesses to respond to reviews; it’s a bit different from what Citysearch is doing.)
While Yelp has a strong, engaged community it becomes relatively easy in selected categories for Google, for example, to duplicate this summary/sentiment analysis function:
So as we get more reviews for more businesses over time, from a consumer perspective, other things start to factor in: my network’s opinion or accessing word of mouth for very specific recommendations.
To varying degrees Facebook, GoodRec, Aardvark, Twitter (to some degree) and the new AlikeList contemplate this less anonymous, more “trusted recommendations” scenario. Yelp too has been connecting people on its site. And Citysearch integrated Facebook Connect some time ago, so that I could see who among my friends was on the site.
The reason that Yelp didn’t show me my “friend’s reviews” or allow “sort by friends” is because there weren’t enough reviews to make that functionality meaningful. Here’s my paraphrase of Yelp CEO Jeremy Stoppelman’s previous explanation of why they didn’t initially pursue this approach:
He reminded me that very few people actually create content, despite the popularity of reviews, which typically means that your friends won’t have written anything about most of the businesses you’re interested in. He also said that services that ping your network when you have a request are also going to be challenging because some people aren’t going to want to receive emails all the time (see Facebook fatigue).
AlikeList is betting that it has removed the barriers to content creation and will be able to build enough usage and community to make this “trusted recommendations” functionality meaningful.
We may not be there but we eventually will be entering a period where reviews about everything are everywhere and people look to “secondary” tools (top lists) or analysis (ratings summaries) to make decisions. Comparable to Yelp’s ratings summary feature, I always point to Amazon’s “most helpful/most critical” review capsule in this discussion:
This is the future to some degree. Reviews will always be important (even essential) but they become less of a differentiator over time. Here’s where Grayboxx’s methodology might be relevant; however the execution was very awkward and the site was probably too early.
Some quick final thoughts:
- More reviews across categories means Google can crawl, capture and present these summaries (bad news for competitors)
- Lists and tools that get me to the “best,” “most popular,” “most recommended” become essential as a first step; then I drill down into reviews if I need more detail and information
- Q&A tools and connecting actual people or being able to query the community become more important as well (Yelp, Yahoo Local and Yellow Pages Group do have this today; working to varying degrees)
- Other site content and functionality become important to complement reviews
- Brand and trust matter (getting reviews from Yelp might be preferable to getting them from a less well-known site that offers reviews and shows up in Google results for the same query)
What do you think about all this and the basic question of whether reviews are moving toward “commodity status”?
Screenwerk 
December 23, 2009 at 4:08 pm
More reviews doesn’t make things worse if there is an ability to sort the reviews to get at the highest quality stuff. Amazon’s shown how this can be done. Google is another example: it doesn’t matter how many pages Google finds (100 or 1 million) as long as the most relevant ones show up on the first page.
December 23, 2009 at 4:47 pm
(Found this latest article off of Twitter)
There are additional considerations beyond those you have referenced, Greg.
1. Its quite possible, if not probable that Google has used and is using the Volume of Reviews for Ranking businesses in their 7pac. More currently as they have added sentiment analysis, which you pointed out, Google might weigh sentiment analysis with regard to ranking within the 7 pac.
2. That Google Maps ranking is important. Take all the review sites, aggregate them, and they may not generate the traffic that search in Google probably does for the various review topics and the businesses referenced.
3. For various businesses rampant reviewing is “Dicey”. Within Google Maps and showing in the 7pac in aggregate volume…then within the local business record…there are an increasing volume of “spam reviews” or attack reviews. One business generates “fake reviews” in maps.google.com and plants them on another site. In some sites you then see “Defense Reviews”. The attacked business generates reviews claiming the others are attacks.
How ugly, how juvenile, and ultimately how much does it affect the visitor to the site and business?
4. As has been pointed out within Mike’s blog; http://www.blumenthals.com/blog by some commentators some industries don’t tend to generate reviews…but then a couple of businesses within those industries “plant” or use enhanced communications tools to generate reviews. The volume of reviews for a couple of businesses gets out of whack relative to all the rest.
What is one to make of that?
There needs to be some enhancements on reviews. Yelp’s summaries are nice, IMHO. I actually liked that most recent 30 day average of reviews in Yelp. A restaurant can change personnel, management, or ownership and it can change very very dramatically. AlikeList ‘s enhancements are nice.
There are pros and cons to the entire review process. If I’m buying on EBay I want reviews on the seller. Invaluable. If I’m reading reviews on a dentist or doctor and there is a mixed bag of commentary…I really want to know the opinion of someone I trust.
Its a topic to continue to watch.
December 23, 2009 at 5:06 pm
Great article/topic Greg.
At a high level I would say too many reviews does start to make these pages more difficult to derive any conclusions from. As more and more people review you are bound to get more and more “one-offs”. I remember looking on Yelp for an Pizza in Phoenix and reading through a ton of reviews, mostly positive, and in the end agreeing with the few negative reviews. The challenge for these sites is how to manage data (if you’re lucky enough to have it) and make it relevant for the user without seemingly hiding data. I think for Yelp, given there business model, will have trouble trying to claim to not to be bias if they start to filter out reviews or make specific recommendations.
I think breaking out the local world into discrete use cases and personas helps to see that there is (obviously) a huge market for sites like Yelp but many other services that have a place as well. Some people like to get advice from friends (alikelist seems founded on this) while others (like myself) have saturated my friends recommendations and am interested in recommendations based on personal preferences/tastes.
One thing I’ve been thinking about recently is how different the use case for local search versus local discovery is…
December 23, 2009 at 9:01 pm
Thanks Dave, Frank. This is obviously a complex area. Agree that there’s lots of innovation in this area that’s possible: recommendations based on patterns/preferences, long-tail discovery, and other types of content/color that aren’t explicitly about reviews.
December 23, 2009 at 9:12 pm
To the topic of reviews/ratings as commodities, it’s interesting that none of the major sites (at least that I’m aware of) allow you to export/import reviews. Our site struggles from getting enough data to be able to mine it for patterns, etc to do intelligent recommendations. Some sites have APIs but I can’t think of any that allow a user to take his or her ratings and allow them to be used by other sites.
January 5, 2010 at 7:07 pm
Greg,
Why do you think Yelp (and other similar sites) haven’t been able to gain enough traction in some verticals like Healthcare (Doctors)? It seems they keep gaining mindshare around restaurants and some services, but haven’t much headway into other areas.
January 6, 2010 at 4:09 pm
Mike:
If you look on Yelp there are dentist and doctor reviews, just not quite at the same volume level. According to COO Geoff Donaker on Yelp, 29% of reviews are restaurants, 22% have a “retail” focus
January 11, 2010 at 10:34 am
Ran into this while I was staying in SF recently. SF has tons of reviews on Yelp and within one area, Sunset, there were dozens of restaurants with roughly 4 star averages. What I needed was 1 restaurant. Perhaps a simple ranking of restaurants in a visually neat, compact fashion would simplify my experience on Yelp; after all, I’m goal-driven in search of a recommendation not on an information retrieval mission. The Amazon feature of ranking reviews described in the blog post is along the same lines- actionable information up front (if you want to research more and read other review, you can).
February 16, 2010 at 1:56 am
Reviews are great as long as they are sorted by relevance. People like reviews, it gives them a nice, warm feeling inside.
September 18, 2013 at 2:07 am
I think as long as the testimonials are real patients that said ‘real’ things about the quality of their exeperience then fine. Howeever, if they go the ‘info-mercial’ route, then no way!