Silicon Valley is a bubble world: GigaOm on the Google/Yelp deal rumors

Monday, 21 Dec 2009

Exhibit A

I am rather confused by how anyone could rationally think that the reason that Google ought not buy Yelp is because Twitter and FourSquare will replace Yelp.

1. a website like FourSquare… that had 60000 users in mid-October (according to the NYT) is far from an established success story. Just because it was covered favorably by the technorati in their blogs does not a game changer make. Far too often hype is conflated with actual success in Silicon Valley… and this is a perfect example. This is a new idea that has had limited traction in 9 months. Yelp has little reason to be scared – just as Facebook has successfully incorporated many features from Twitter, Yelp can incorporate the good product ideas from FourSquare. There’s the “Talk section” on Yelp already – this is a section of the site that would blossom if incorporated into the Yelp mobile apps to give users real-time feedback on what’s hot and fun nearby.
2. Having real-time feedback is only useful if I have an extensive online network of friends. What if I am in a city where I know maybe one or two friends? How do I get real-time feedback? What are the odds those couple friends will be online and checking their Twitter EXACTLY WHEN I AM LOOKING FOR A BITE TO EAT? At best, Om’s argument means that FourSquare and Twitter would work as real-time review feedback tools for places you live or have lived in the past (and thereby know a huge concentration of people who go to restaurants and bars in the neighborhoods you hang out in).
3. One possible counterpoint to this argument is that: in the future, people will have way more FourSquare and Twitter friends, thereby mitigating this problem since for most cities, people will have a trusted source of friends from which to obtain real-time feedback for where to eat/drink/see shows/whatever. This leaves Om’s position in a double-bind. As users extend their networks, the commonalities between the users and those that they add to their network decreases. To put it another way: if I follow 10000 people on Twitter, the likelihood that my restaurant/bar/whatever choices align with these 10000 people decreases, as compared to when I had a network of 100 close friends who I followed. This reveals a fundamental tension: in order for real-time feedback to be a real asset, a user’s network must extend to huge proportions. But the value of this feedback, according to Om, is that it “involves my social graph”… a social graph that is made increasingly superficial and meaningless as my network expands to cover all the cities in which I want to find a bar or restaurant. That leaves two options, both of which do not leave me convinced that Twitter/FourSquare can do anything to supplant Yelp’s market position: either lots of meaningless data or little tiny bits of meaningful data.
4. A lot of my friends have really shitty taste in food. Some of them like going to trashy bars full of douchebags. Just because I trust a friend for one thing (taking me to bars where I will have a great time) does not mean I trust them for another thing (telling me a great date restaurant). If I want personal advice from within my friend network, it is probably easier to write an email or call specific friends rather than solicit advice from hundreds of people who I may have arbitrarily added online on some website.
5. Om’s article assumes that there is an inherent value to real-time information. This is true for breaking news; it is true for customer service responses over Twitter. It is 100% false for reviews of businesses. In fact, having a chronological history of reviews available to me provides useful information about the establishment that help me make decisions – i.e. they may have changed chefs or owners recently, etc. I can probably safely disregard a lot of reviews for a restaurant’s food if the chefs and menu changed 3 months ago; I have no idea if my friend who is providing real-time feedback in response to my tweet went here 3 or 6 months ago, and the 140-character limit on tweets just decreases the likelihood that I get any kind of useful background information from the reviewer. Having a history of reviews by a reviewer is also important – for example, the average Bay Area resident loves the [fake] NY-style pizza served at places like Amici’s and Escape from NY. I almost universally hate all pizza I have had in the Bay Area, with very few exceptions (pizza that doesn’t hit the spot, but is OK and fulfills my pizza cravings), so a Bay Area resident would be very reasonable to ignore my reviews of pizza joints in the area (none of them get over a 3), or at the very least weigh my reviews differently in his/her decision calculus about where to grab a slice.
6. There are plenty of other problems with Yelp’s business model, as well as their advertising sales tactics (I am sure my negative 1 star reviews have generated Yelp tons of money… maybe I can get a referral for helping them generate leads to call businesses I think suck?). In fact, I could think of a number of reasons why Google may not want to do this deal that have nothing to do with Yelp being replaced by FourSquare/Twitter.

Update:
“Couldn’t agree more. The main reason Twitter + Foursquare trumps Yelp is simplicity. I don’t have the time nor the patience (unlike the “Elite”) to review every restaurant I visit. It’s either 140 characters or Zero characters worth my time ” – from the first comment on the article.

I cannot think of many things I can communicate in 140 characters or less that are in any way informative. In fact, if I go to Twitter.com and click on a random hashtag that is a “trending topic”, 90% of the tweet results are uTtEr JiBB3r4sH (translation: utter jibberish) written like so; maybe the 140 character limit has inspired a modern literary masterpiece (post-modern haiku?) for the ADD generation, but I certainly have not discovered anything remotely informative on Twitter other than people reposting links to cool articles that are not actually written in a 140 character or less blurb.