Photo: Facebook Search
Yesterday, Facebook finally entered the search business.Ostensibly, this is great news for Facebook, users, shareholders, customers, and employees, as “search” is one of the most useful and lucrative applications in the digital world.
Ever since Mark Zuckerberg took the stage at a conference in November with the stock at $18 and said, yes, of course Facebook would go into search at some point, investors have been on the edge of their seats.
And excitement about that new revenue engine, as well as expected acceleration in the core business thanks to Facebook’s new mobile products, has driven the stock back up to $30.
It is indeed exciting and promising that Facebook has entered this business.
But the truth is that Facebook’s current beta product is miles from being anything more than a cool, revenue-less feature that some Facebook users might want to play with sometimes.
It isn’t yet something that is going to be particularly useful.
And, importantly, it isn’t yet something that will compete with Google for search ad dollars.
Our West Coast editor, Owen Thomas, took Facebook’s new search engine for a spin yesterday.
As you can see here, it’s a cool way to search Facebook that takes advantage of the fact that Facebook is a big structured database. And if you’re looking for pictures that people have uploaded to Facebook (and made public), or friends who went to a particular school or live in a particular place, or potential mates, it seems quite useful.
Beyond that, though, the search engine indexes too little data to be valuable as a commercial discovery engine.
One of the immediate commercial uses that everyone began hyperventilating about yesterday, for example, was the hypothetical ability to use Facebook Search to find great restaurants. Everyone would just type in “Indian restaurants my friends like in San Francisco,” this theory went, and they’d immediately find the good ones. Yelp would be toast, and all restaurants would immediately start paying Facebook gobs of money for sponsored listings, and so forth.
Well, Owen tried a few restaurant queries, too.
The results were, to use his word, “awful.“
Why were they awful?
Because Facebook doesn’t have enough data about restaurants that Owen’s friends like to be truly useful. To have enough data about those restaurants, Facebook would want to go out to the web–to Yelp, to Google, to Zagat’s–and begin to incorporate other factors despite just “likes.” And Facebook would have to add detailed comments, and pictures, and ratings–the sorts of other structured data that anyone trying a new restaurant would want to know. And Owen’s friends would have to eat at, and “like,” more restaurants.
(It would be nice to think that folks will happily rush to a restaurant because a couple of friends have “liked” it on Facebook, but most Facebook users have used Facebook enough to know that “liking” something doesn’t always mean that you actually liked it. And just knowing that someone “liked” something would leave most people wanting more information about it. Furthermore, “likes” will become even less meaningful if Facebook’s new search engine suddenly prompts even more companies to pay to get “liked,” which is one possible side-effect of the search engine that our Advertising editor Jim Edwards wrote about.)
More importantly, Facebook’s new search engine has not yet entered the truly lucrative portion of the search business, which is standard online commerce.
The truth about Google is that only a small portion of Google search queries produce any revenue. And the ones that produce revenue are generally commerce queries: People looking for specific products to buy.
That’s the search inventory that advertisers go bananas over.
That’s the “database of purchase intentions” that has powered Google to a ~$50 billion business.
Right now, if you type a typical commerce search query into Facebook’s search engine (I’m sorry, but I just can’t bring myself to call it by its terrible name, “Graph Search,” because no one outside the techo-chamber has any idea what that means), nothing happens.
For example, if you type in “Digital cameras that my friends like,” you don’t get jack.
To make Facebook Search useful for that sort of query, therefore, Facebook is going to have to continue to add capabilities to the product.
And eventually–and this is the critical step–Facebook is going to have to expand Facebook search to the web.
Not just the type and number of products that are indexed in Facebook Search.
But the actual search window itself.
As Microsoft and many others have made clear, the search business is a scale business: To make your query results relevant, you need a vast number of queries. And to get your search business running well, at a high revenue per search, you also need a vast number of queries–because it is only with this vast number of queries that you can build enough intelligence into the system that you can deliver the right sponsored ad at the right time and, thereby, drive up your revenue and the advertiser’s ROI.
Google is exceptional at doing that… in part because Google has ~65% of the desktop search market (vastly more, if you include mobile and international)
Microsoft has about ~25% of the desktop search market, and this isn’t anywhere near enough for Microsoft to make money on search. Microsoft can generate search revenue, but not profit: The amount it spends to generate its queries and traffic more than offsets the lower revenue-per-search that it can generate.
Facebook, meanwhile, only has about 2% of U.S. based desktop query share.
And the number of commercial queries in Facebook’s 2% of queries is likely even proportionally lower.
So, to make this a good business–in addition to a truly useful commercial tool–Facebook is going to have to:
- Expand its search database to the web (thus capturing vastly more “likes”)
- Generate results that include input from more than justyour friends (to get a meaningul way of distinguishing between, say, a couple of local restaurants, Facebook is going to have to include “like” data from hundreds of thousands of people who fit your demographic and general tastes and so forth–your few hundred friends just aren’t a big enough sample. And your friends might not like the same food you do. Etc.)
- Move its search window and service out onto the web, en masse, and into “browsers” and other gateways. If Facebook ever wants to capture real search dollars and present truly relevant results, Facebook will need to radically increase the number of commercial search queries it is processing. And, right now, most of those queries take place off of Facebook.
Bottom line, Facebook Search is a cool new Facebook feature. Lots of people will probably play with it, at least initially–just the way they played with “Siri,” when Apple rolled Siri out.
(And, as a reminder, when Apple demoed Siri, lots of people said it was going to kill Google. That idea is laughable now.)
But for this to be a truly useful product–and, more importantly, a real business–Facebook is going to have to do much more (and much differently) than it has done to date.
In other words, the hope that “search” will suddenly radically accelerate Facebook’s revenue growth and deliver on the financial hopes and dreams that investors had about the company a couple of years ago remains just that, hope.
There’s a lot of potential here.
But Facebook Search has a long, long way to go before it will ever really matter to anyone.