Part of what we do at allLocal is help national and regional businesses understand how their brand is viewed at the local level. Based on this work, we have developed a rich set of data around the local review space and we thought it was time to shape this data into a bit of retrospective on the space. The earliest local review we have analyzed for our customers is from May 2002 and comes from Citysearch. In the almost 9 years since this review was written the local review space has grown…but has not really changed. The only real innovation in the space is the community / vanity aspect that Yelp brought to the table in 2004. While the number of reviews and the pervasiveness of them in the local search experience has grown, there has been no dramatic shift in the space. As for the future, I think our best bet lies with the social networks and their ability to passively collect this information as part of a conversation as opposed to the explicit ‘user writes a review’ model that we have today. After all, local reviews are so valuable because they are taking information that typically does not exist in a digital form and digitizing it…making it crunchable by an algorithm. Who says this digitization has to be done actively and can’t just be something that happens as a user lives their digital life? Enough prognosticating though, let’s get down to the data:
About The Data
- ~2,200 locations distributed across 4 general verticals: 45% travel, 40% retail, 13% services and 2% restaurants
- Analyzed the number of reviews written at Yelp, Google, Yahoo, Citysearch and Yellowpages.com. We specifically focused on review sites that intend to have a broad appeal and because of this we left sites like TripAdvisor out of the mix. We will evaluate them in the future when we take a more vertical specific look at reviews
- For each engine we looked at the number of reviews written for any of the 2,200 locations on each day between January 1, 2005 and January 30, 2011. The numbers you see tracked here are the 14 day moving average of the total number of reviews we saw written each day, divided by 2,200 to get the daily reviews per location and then multiplied by 100 to get a more readable number. Basically, we took the 14 day moving average of this: New Reviews / 2,200 * 100
The 5 Year Retrospective
Some observations based on this:
- The number of new reviews created each day across the 5 engines as a whole has grown by ~10x over the last 6 years
- Yahoo! had the early lead in the space but did not capitalize on it (I was able to copy and paste that line from other articles that I have mentioned Yahoo! in…)
- Yelp had a slow burn in the beginning and then really broke out in the later half of 2008
The 2 Year Retrospective
When we zoom into a 2 year range, some other interesting observations emerge:
- The trend is Yelp’s friend
- Google permanently passed Yahoo! based on new local review growth in early 2010
- Citysearch and Yellowpages continue to struggle in getting user participation on their sites (i.e., getting people to write reviews)
- The pop in Google reviews in the middle of November 2010 is interesting…
The 8 Month Retrospective
Looking at the last 8 months give us a few more things to consider:
- In a relative sense, Citysearch had a good holiday season and generated a bunch of new reviews from their users…though they are still far behind in an absolute sense
- Yahoo! had two peaks in review activity in late 2010…and based on a quick review the bursts seems to be very methodical and almost robotic in nature. Activity like this stands out when you are dealing with what is still a relatively small data set at 2,200 locations
- That Google pop is worth discussing further…so let’s add some commentary to the graph
Ahh yes, Google Hotpot went live on November 15th. Looks like they got a nice pop from the launch but then quickly settled back into their (flat) trend line. Ouch….but to be fair, this analysis is not restaurant centric and Hotpot, while available across all types of business, is certainly angled towards the restaurant industry. But then the same could be said for Yelp. And speaking of Yelp, their trend line either got a nice boost from the holidays or from the Hotpot launch reminding everyone to go write some more reviews on Yelp.
This trend needs to be watched as the success of Hotpot is critical to Google. Their review aggregation approach has crumbled under the weight of Yelp and now TripAdvisor not always playing nice with Google. However, even if they do play nice a review scraped off some 3rd party site will never provide the insights that a review written directly on Google will. All the meta-data that comes along with a direct review (i.e., the history of the person who wrote the review) allows Google to place much more (or less) trust on the authenticity of the review. For scraped reviews, Google either needs to rely on the 3rd party sites to police their reviews effectively or has to make the decision themselves based on much less data. We continue to see the importance of review text rise in Google Places ranking but access to fresh and trusted reviews is needed to make that model work. Without Hotpot, Google may lose the local digitization race which, in my opinion, will ultimately decide who wins the local search race.
Thanks for taking the time to review some local reviews with us. Lots more to look at in this space and our next post about reviews will dive into either a vertical specific or a geographic specific look at the review space as both provide some interesting data.