Where Do My Friends Fit Into My Local Search?
A longstanding question has re-emerged lately in the tech blogosphere: Do direct social connections add value to finding things to do, see, or buy locally?
In other words, do recommendations for a contractor, flat screen TV, or where to eat lunch carry greater weight from my friends than the larger — albeit less intimate — wisdom of the crowds. For the former, think Facebook friends. For the latter, think hundreds of Yelp reviews.
The question’s importance accelerated after Facebook first launched Graph Search, and the internet went crazy that the latest “Google Killer” had arrived. Instead of relevance governed by keyword and page prominence, social data (likes, etc.) were thought to rule as ranking factors.
For example, when visiting Cleveland, I might research pubs that align with my affinity for craft beer. Searching Facebook for “Cleveland gastropubs” would hopefully surface related posts or comments from friends (or friends of friends), in addition to all Facebook users.
This is basically the digital version of the word-of-mouth factor that’s always been a leading source of influence for local commerce. Therefore the common wisdom in local has been that any search or discovery tool can be vaulted by infusing sentiments of the social graph.
I mostly agree with this thesis, but have begun to amend my position: Scale trumps peer recommendations in influencing buying decisions. That’s because local search is a game won on a combination of relevance and volume. The social graph brings relevance (trust, etc.), but not as much volume.
For example, Facebook is limited to its own walled garden. Though it’s a pretty big garden (1.5 billion users), it’s dwarfed by Google’s three trillion-page index from which to answer long-tail queries. What contractors work with colonial design? What pubs serve Pliny the Elder?
Back to the Cleveland example: will my 789 Facebook friends have enough relevant sentiment on the beer scene there? Google’s index (or Yelp) will unearth a more statistically significant sentiment. The same goes for any nuanced query — such as highly customized home services.
Similarly, there’s an argument that questions the reliability of 5000 Yelp or Google reviews, in favor of 4 or 5 recommendations from my “trusted” friends. This also doesn’t hold up due to one key factor: I’m friends with people for reasons other than a shared taste in food.
In other words, the social connections that bind us don’t correlate to food taste or flat screen TV quality. These are things for which sample size trumps personal connections in surfacing the truth about a product. There are exceptions where trust is central (i.e. child care).
This is all to say that Google’s strengths in local lie partly in the scale of its index. Even as the entry points to local discovery fragment into apps instead of search, those apps individually lack the utility of a massive index — even Facebook, with all of its mobile reach, falls short.
So Google’s goal in adapting to a mobile world is to index and unearth content from disparate apps, just as it now does for the web. This can be seen in what it’s currently doing with Now on Tap, deep linking, the “hub and spoke” approach, and other initiatives yet to be revealed.
Meanwhile, social graph sentiment can be pulled into a hub like Google Now. That data could indeed add dimension to local discovery – similar to what Google already does with reviews in GMB listings. In a related move, Google recently pulled tweets into mobile search results.
But the social graph alone won’t be a silver bullet for local. It’s just not big enough. Three years into Graph Search for instance, social’s value to local is still overstated. My friends are great for party pics and snappy news feed dialogue — they don’t yet excel at plumber reviews.
Michael Boland is chief analyst and VP of content at BIA/Kelsey. Previously, he was a tech journalist for Forbes, Red Herring, Business 2.0, and other outlets.