No. 751
December 15th, 2010

Hotpot vs Yelp vs Facebook

Obviously, it’s been a whirlwind last couple of months in Local Search…it may take all of next week to write the 2010 version of the Year in Review.  But I’ve had a few ideas bouncing around in my brain about Hotpot recently–why Google released it, and why they’re promoting it so hard, to the point of breaking their own rules on Review Solicitation.*

*Based on the follow-up statement they gave Matt, I’d expect their guidelines to be updated shortly to include the word ‘positive’ in this sentence: “In addition, we do not accept reviews written for money or other incentives.”

HotPot Addresses Google’s Two Biggest Local Vulnerabilities

Google has completed an amazing technological feat with Places.  It gets searchers directly in touch with local business information, rather than sending searchers through an additional layer of search results, more quickly than any other site on teh Interwebz. Yes, there have been plenty of merging and mapspam issues along the way, but the current product does a better job of “organizing the world’s” location “information” better than any competitor.

What are Google’s two biggest Local vulnerabilities? Lack of “in-house” user reviews and the lack of a social graph.

Hearing Places Product Manager Carter Maslan’s ILM ’10 presentation last week, it’s clear that his focus is all about surfacing the best Local information no matter where it comes from. I’d argue that the “best local information” currently lies outside of Google’s oeuvre.  Beyond basic business information and links to the business’s own website, I’m not sure people naturally think of Google to answer the questions “I wonder what other people say about this business?” or “I wonder what my friends would recommend?”  Its laser-like algorithmic focus becomes a double-edged sword, as people use Google to get the information they’re looking for quickly, without “giving back” to other searchers what their experience of that business was actually like.

As I mentioned in my GOOG-Pon column a couple weeks ago, one of the things that I think Google found so attractive about Groupon was the potential for an insane review count–30+ million if each Groupon user left even one review about their experience at a business they’d visited.  Carter was tight-lipped about how many native Places reviews Google had, but compare that 30 million number to the 14 million Yelp reviews Jeremy Stoppelman was beaming about last week (also at the Kelsey show).

As more and more local sites (including Yelp, based on Stoppelman’s tone in Santa Clara) become antagonistic towards Google’s methods of surfacing their content, Google is clearly looking for more content submitted by Google users directly.  One of the main hurdles to acquiring this content that I think Google has faced to date is the difficulty for searchers to leave a review.  HotPot, though, encourages a persistent login to receive its benefits (see more in bullet #3 below) and leaving a review (or perhaps more accurately a rating) is now a one-click process for HotPot users.

Then, of course, there’s the social angle.  Hotpot lies somewhere between uber-geeky Location-Based-Services like Foursquare and Gowalla and uber-mainstream Facebook.  Although Facebook has not executed well (at all?) on its Places strategy yet, I think Googlers saw its incredible potential as soon as Zuckerberg announced it.  Given Google’s algorithmic track record, and its vast institutional experience in Local already, I’m betting its engineers will be able to come up with a more targeted social recommendation algorithm, and be able to do so more quickly than Facebook.  They just need a critical mass of users to adopt this new technology.

HotPot Puts the Essence of Yelp to the Test

Despite being known as the most comprehensive source of local business reviews on the web, Yelp has been strangely adamant about business owners refraining from soliciting reviews from Yelpers.  The motive behind this stance has always been clear: Yelp feels that the very act of encouraging a review erodes the trust level that its users will have in the veracity of those reviews.

They’ve also been vehemently opposed to quick-hitting ratings from individual users, intentionally crippling the ability for mobile searchers and app users to leave reviews directly from their iPhones or Blackberries.  Yelp has actively cultivated deep tomes of content about what the full experience of a business is like, building in a rewards system for community members who leave the most humorous, most detailed, or most insightful reviews.

Perhaps HotPot will introduce these kinds of badges and leaderboard systems eventually.  But Google is clearly coming at this from a different angle.  They appear to be far more interested in the broadest possible array of quick-hitting, numerically-based ratings. Where Yelp is all about depth, HotPot is all about width, in terms of trying to glean at least a modicum of rating information about each business from as many of its patrons as possible.  Focusing on these discrete signals (as opposed to trying to analyze and assimilate disparate textual sentiments) makes their job of recommending successful businesses to Google users much simpler.

Clearly the effort it takes to leave a super-deep review on Yelp provides somewhat of a built-in anti-spam mechanism.  Google may have a tough time combating automated or quasi-automated HotPot abuse.

But the experience on a mobile device is all about knowing where to go quickly, to use Stoppelman’s words, while you’re “on-the-go.”  My own personal experience, even on a desktop, is that I don’t have time, energy, or interest to read paragraphs, let alone chapters, of content about a place I want to go to dinner.  Star ratings from a reasonably-sized cross section of customers are usually good enough for me.

Yelp will maintain an incredibly devout following among who enjoy the more social and creative experience of sharing their recommendations, but as the broader population gets more and more addicted to its smartphones, it will be interesting to watch whether Stoppelman & Co. adjust their strategy in some way to start to attract a wider base of ratings.

HotPot Has Significant Potential to Addict Searchers to Google Even More

When HotPot first launched, I mistakenly looked at it solely from the data angle–as a way purely for Google to get a broader base of reliable algorithmic signals with which to rate local businesses, and as a preemptive effort to get into the social game before Facebook Places became fully-baked.

But, speaking to several Local Search industry folks in the past 10 days, I’ve realized what many of them have told me: HotPot has the potential to be so much more…in short, the Netflix of Local Search.

Netflix famously offers millions of dollars in prize money for companies or individuals who can improve on its recommendation algorithm.  Why?  To keep users addicted to its service as they browse (aka search) movie titles in its database.  In much the same way, I think Google views HotPot as a way to get (or continue to be?) way out in front of everyone else in the Local Search space from an algorithmic standpoint and keep its users drinking the proverbial Kool-aid.

Leveraging the greatest collective algorithmic genius in the history of human civilization, HotPot has the potential to be far more successful than Facebook Places by combining the ratings of other similar users with all of Google’s traditional Places algorithmic signals like review volume, review velocity, link authority, geosocial signals, and–maybe most importantly for HotPot–sentiment analysis.  No one else is going to have that incredible combination.

Facebook Places is going to have a tougher time, in my opinion, because at an algorithmic level, it will always be harder to judge what kinds of friends you actually want to receive recommendations from on a category-by-category basis.  By encouraging users to form as many relationships as possible with other Facebook users, Facebook has polluted its own recommendation pool to some extent.  And even among close friends, tastes can vary widely.

While I might value my friend Marshall Simmonds‘ recommendations for microbreweries or sushi restaurants, I happen to know that his taste in golf courses is absolutely terrible, and seeing an endorsement from him would be the kiss of death for any course in my little black book.  Since Facebook is likely to rely more strongly on recommendations from within a person’s social circle, I think they’re going to have a much more difficult time sorting out the exact nature of those relationships, and in what categories I’m going to want recommendations from whom.  Whereas Google’s cross-sectional, behaviorist approach to recommending local businesses is likely to result in fewer false positives or undeserved negatives from one’s social circle.

Whatever HotPot eventually morphs into, or what market share it winds up accruing, it’s clear that review acquisition will continue to become an even more critical piece of search engine optimization.  Every customer who walks in the door will have the potential to be a social maven or HotPot everyman.