Wednesday, November 21, 2012

New Search Engine : Yahoo and Facebook Will Team Up Together


The world, it seems, just cannot wait for Facebook to come along and start its own search engine and knock Google off the top of search mountain. The rumors and expectations of this go back for years. Now the latest twist: Yahoo and Facebook will team up together. If you’re banking on that potential alliance taking out Google, much less Bing, let me offer up a reality check.

Report: Yahoo & Facebook Talking Search
Let’s start with the current news that has various tech blogs lighting up over the possibilities.

The Telegraph, citing anonymous sources, reports that Yahoo CEO Marissa Mayer and Facebook COO Sheryl Sandberg have had talks about working more closely together, in particular on web search. From the report:

Sources have told The Sunday Telegraph that Marissa Mayer, chief executive of Yahoo!, has held discussions with Sheryl Sandberg, Facebook’s chief operating officer, about how the two companies can work more closely together.
The two internet giants have already collaborated together on a number of small projects, for example to share Yahoo! news on Facebook, and recently agreed to settle a number of long-standing lawsuits over patents. However, board members expect the talks to lead to much more substantial collaboration based around web-based search.
That’s all the substance there is to this report. Yahoo board members, it seems, believe that Facebook will want to work with Yahoo on “web-based search.”

Debunking The Rosy Assumptions
The rest of the article is full of speculative assumptions, perhaps things fed to the reporter by those sources.

On the face of it, these seem to make a compelling argument for a Yahoo-Facebook search partnership. But, if you know anything about the web search space, they quickly fall apart. Debunking time.

Facebook Already Has A Search Partner: Bing
First up, the idea from the article that Yahoo somehow has something that will help Facebook excel with search:

Forging an alliance with Yahoo! would allow the social network to take a major leap forward in search, ensuring it remains central to its users’ lives and helping to target advertising more efficiently.
Really? Why? Why would an alliance with Yahoo do something for Facebook that Facebook’s existing partnership with Bing doesn’t already do?

Remember? The partnership that kicked off in October 2010, when Bing started integrating Facebook’s social data into its results? When Facebook CEO Mark Zuckerberg said this about working with Bing:

“I couldn’t think of anyone better to be working with to build the next generation of search.”
Only two years later, and what, Bing’s not been working hard to deliver on that next generation? I guess Facebook wasn’t happy with that whole redesign Bing just did, with a social sidebar inviting everyone who searches on Bing to also share with friends on Facebook:

The answer is there’s nothing Yahoo brings to the table that Facebook doesn’t already get with Bing. Bing has far more search talent than Yahoo. It has far better search technology than Yahoo. At this point, Bing even has more useful web search experience than Yahoo, given that Yahoo has outsourced web search to Bing since 2010.

The Mythology Of Scale
Next we get this:
Meanwhile, Yahoo!, which started life as one the first major search engines but is now dwarfed by rival Google, would benefit from Facebook’s vast army of more than 1bn users.
The computer codes which power search engines become more powerful as more people use them, making it tough for Yahoo! to stage a comeback without help from another, more popular organisation.
The computer codes that power search engines — algorithms — don’t become more powerful as more people use them. They become more powerful when smart people program them better. If the programs just naturally got smarter with a lot of use, Yahoo would still be the leader in search and Google would have never emerged.

Now, it can definitely help smart people to program better algorithms if they have lots of real-life data to model. That’s where this whole “scale” aspect comes in, something in particular that Yahoo and Microsoft tossed around a lot to sell people on the idea in 2009 that if those two partnered together, they’d have enough data to finally beat Google.

The thing is, having a 10% share of all the searches in the United States is probably plenty of data to model on. It’s a huge amount. Yahoo still has more than that already, 12%, according to the most recent comScore stats. Yahoo potentially sees more than this as part of its deal with Bing. Yahoo also had more than this back when it went down the road of walking hand-in-hand with Microsoft.

Add to this the fact that the tiny Blekko search engine seems to be churning out improved search results despite having a share of search so small that comScore doesn’t even report it. It’s not that 10% of searches is enough. Even 1% of the searches in the United States, or perhaps 0.5% of the searches, might still be plenty of data to help you improve your search engine. Remember, Google somehow managed to grow from no share to where it’s at now. Search data helps, but lots of search data isn’t some magic solution.

Given this, no, Yahoo doesn’t desperately need a horde of Facebook users to fill the gap — users, by the way, who Facebook itself says primarily are searching on Facebook other people on Facebook. That’s not a rich source of data to mine for improving web search quality.

Facebook Will Make Working At Yahoo Awesome!
Next, a “Facebook Factor” is suggested as somehow pumping up Yahoo:

Working with Facebook would also allow Yahoo! to piggyback on the social network’s brand cachet to help it recruit top-tier computer programmers – something that has been a major problem for Yahoo! in the past two years as it has hired and fired a string of chief executives.
Let’s get this straight. If Yahoo partners with Facebook, that will help it attract computer programmers — presumably search engineers — because of some “Facebook is cool, so we are too” factor?

If that’s the case, it should work even better for Bing. If you’re a search engineer, and you want to work with a search company because it partners with the cool kids over at Facebook, you can go work for Bing now rather than hope that maybe some Yahoo-Facebook deal will happen that maybe will result in a revived Yahoo search engine.

And yes, Bing even has a Silicon Valley location, if being in the Bingplex in Bellevue, Washington is just too cloudy for you (and it’s not that cloudy). Until recently, it was even run by Sean Suchter, who came to Bing from a little company called Yahoo.

Yahoo The Kingmaker? The Googleslayer?
So far, the things I’ve debunked seem to be coming from the Telegraph reporter, but this next part suggests that it might be a Yahoo person (perhaps a Yahoo board member, I’m guessing, and one almost certainly way too optimistic), painting a picture of how great all this would be:

An alliance between Facebook and Yahoo! will pose a major threat to Google and stands to reorder the hierarchy of the world’s biggest technology companies. A senior figure likened Yahoo!’s position to that of a minority party in a hung parliament, with the power to act as kingmaker by choosing another party with which to align itself.
Yahoo the Kingmaker? Yahoo, with a search share that’s in decline, with no real search technology, is somehow going to magically transform Facebook into something that will pose a major threat to Google, presumably in search?

It’s hard not to laugh. Really. What next? AOL will prove crucial to help revive Microsoft’s fortunes in the mobile space?

Yahoo May Abandon Microsoft (Yep, But Not For Facebook)
Let me cap off the debunking with this:

Silicon Valley observers had speculated that Ms Mayer could throw Yahoo!’s lot in with Google, her former employer, but instead she and Ms Sandberg, who is also a former Google executive, are expected to use their combined might to launch a serious competitor.
However, Google is not the only major technology business under threat from the Facebook-Yahoo! alliance. It also throws Yahoo!’s relationship with Microsoft into doubt.

OK, Yahoo is indeed not happy with Microsoft, and for good reason. Yahoo knowingly entered into a deal with Microsoft that was supposedly going to produce oodles of cash all around but which has proved disappointing.

Why & When Yahoo Might Leave Microsoft
Quarter after quarter, there are promises it will get better. In reality, the “revenue per search gap” hasn’t been closed. If that isn’t fixed, Yahoo has the ability on March 31 of next year to walk away from Microsoft. This all detailed in my article from earlier this year:

As The Yahoo-Microsoft Search Alliance Falls Short, Could A Yahoo-Google Deal Emerge?
That’s a required reading piece, if you really want to understand how fanciful an idea it is that Yahoo is just going to magically walk away from Bing and into the arms of Facebook rather than Google, to solve its search woes.

Yahoo lacks core search technology — not just the ability to index billions of web pages in a timely manner and comb through those with an algorithm to find matching ones — but also to serve up search-related ads in a way that generates as much income as the industry leader Google does.

Indeed, the failure of Yahoo — when it had far more search talent and share than it does now — to make its “Panama” search ad system generate more revenue is a key reason why investors started losing faith in the company, driving it eventually into the arms of Microsoft.

In short, for Yahoo to do search on its own, it effectively has to start over at the beginning, building its own search race car from scratch while Google and Bing keep lapping it around the track. Maybe Yahoo will decide to do that with Facebook’s help, but Facebook really isn’t a help. It doesn’t run its own search engine. It has no experience to offer here.

Why Facebook Can’t Help (Much) With Search
But wait, I hear some saying. Facebook sees a gazillion Likes from all over the the web, and it does gather info from those pages, and it sure knows a lot about scale, and gee-whiz they have a lot of smart people at Facebook!

Look, Facebook is amazing in many ways. Sometimes, the stuff Facebook surfaces in my news feed by analyzing my relationships with people is uncanny. There’s no doubting Facebook’s engineering prowess. But web search is nothing like running a social network. Web search is about:

Gathering up tens of billions of web pages
Gathering up some of those pages within seconds of being published, so you don’t appear stale on breaking topics
Revisiting all your pages in a timely manner, often every couple of days
Figuring out a way to rank those pages to show the most relevant ones
Figuring out a way to rank those pages when terms have multiple meanings, like “jaguar” or “apple” or “football”
Fending off poor relevancy caused by people who, with little effort, can flood the data you collect with millions of crappy pages and bad relevancy signals
That’s just top-of-my-head stuff. Web search is hard. I’ve been covering search now for going on 17 years. I’ve seen all types of search start-ups promise to somehow change the playing field. I’ve seen one actually do it. It was called Google. And Google got there by working very hard, and for years almost single-mindedly, on search.

If Facebook is wanting to challenge Google in search, it would have to make a serious effort, a major engineering investment, and be recruiting plenty of key talent. None of the signs that this is happening are out there. There’s been no major departure of search talent from either Google or Bing that anyone’s noticed — and that would get noticed. It certainly was noticed when Yahoo started losing talent.

Also, if Facebook were going to make such a move, you’d at least expect it to be testing much more heavily what it can do with web search right now, using Bing.

That’s not really happening. Yes, you can get web search results from Bing at Facebook, but you almost have to jump through hoops to get there. That’s the reality. The fantasy is that despite not testing some radical new Facebook Search powered by Bing, its existing and excellent search partner, Facebook is sitting around thinking that Yahoo has the keys to the search kingdom.

Facebook’s Search Challenge
But wait! The Telegraph story notes:

Facebook has already said it plans to boost its web search facility, with founder Mark Zuckerberg noting that the social network is “pretty uniquely positioned to answer the questions people have.”
He sure did say that. I was in the audience at TechCrunch Disrupt when it happened. Here’s my write-up:

Zuckerberg On Search: “At Some Point, We’ll Do It” & Be “Uniquely Positioned”
My take was that Zuckerberg wasn’t saying that Facebook would go head-on against Google in search but rather perhaps might do a better job of showing answers for very popular and personal questions, like if you wanted a sushi restaurant recommendation.

Facebook could certainly do something like that. I’m actually surprised it hasn’t done more in search already. But the answer seems to be that most searches (90% or more, the last time I checked with Facebook) on Facebook continue to be people on Facebook looking for other people on Facebook.

The demand isn’t there to build out a general purpose search engine on Facebook. Facebook users aren’t screaming for it any more than Google users were pleading that Google give them a social network (it did, anyway, and Google+ has hardly been overwhelmed by huge demand).

If Facebook does decide to build a Facebook search engine, then it has a struggle with which of these two boxes is more important:



It’s hard to make one box do two things: be there for search needs or be there for people to provide updates. Facebook, I think wisely, has seen its focus on encouraging the updates that are so crucial to being a social network (and that’s why the Update Status box is more prominent). Search? That’s a potentially valuable revenue source, but is it worth dramatically messing with user habits to try and encourage more of it? Potentially messing with your core revenue stream?

Facebook’s Discovery Engine?
I think Facebook will remain very cautious when it comes to search. I think it’ll poke at the edges where it makes sense, leveraging its search talent more for “discovery,” that is surfacing up answers to needs you might have but haven’t expressed with an actual search.

For example, if you were on a shopping site, and perhaps you were looking at various products, Facebook might be able to know that and suggest related products to you. In fact, that’s already what it does working with retargeting partners, as I’ve covered recently:

Facebook Exchange & How My Drugstore.com Basket Got Turned Into Facebook Ads
But a full-scale search engine? Why? Especially why when Facebook already has a partnership with Bing for high-quality search results?

No Magic Social Signal Solution
As an aside, there are those who think that Facebook has the social data that are the keys to improving search. After all, it knows all the most liked things on the web! But those who’ve assumed a magical social-powered Facebook search engine would emerge back in 2007 to topple Google and again when the Facebook Open Graph launched in 2010 are still waiting.

I actually agree that social data will be very important to improving search results, eventually. But so far, despite Bing having access to all Facebook’s social data, it’s finding it problematic to use it as a replacement for the increasingly terrible ranking signal from links. Google is finding its own social data also difficult to use for the same reason. My article below has more about this:

Links: The Broken “Ballot Box” Used By Google & Bing
Personally, I think the search engines, both Google and Bing, aren’t doing enough to make use of social data. Links weren’t some perfect signal at first. It took years to refine them. I think social signals offer great promise but that the major search engines just aren’t gearing up enough to mine that data.

That also leads to the idea that Google, by not getting any of Facebook’s data, will be left behind. The reality is that even without that data, Google’s already returning top pages that also have strong Facebook Like activity. Not having that data hasn’t harmed its results so far, and Google+ remains an important “ballot box” of its own that can be counted.

Yahoo’s Real Microsoft Alternative: Google
In the end, if Yahoo’s going to leave Microsoft, there’s only one viable candidate to go to: Google. That depends largely on whether it would be allowed to go there by the US Justice Department, which threatened anti-trust action the last time Yahoo and Google talked partnering in 2008.

I don’t doubt that the Telegraph is correct that Yahoo and Facebook are talking about ways to work together, search included. It’s also incredibly intriguing that that one of Google’s former heads of non-paid search, the super-smart Marissa Mayer, is talking with one of Google’s former heads of search ads, the super-smart Sheryl Sandberg. But it’s just difficult to see what it is that Yahoo has to offer to Facebook in terms of search that it doesn’t already get from Bing.

0 comments:

Post a Comment

Twitter Delicious Facebook Digg Stumbleupon Favorites More

 
Design by Free WordPress Themes | Bloggerized by Nilesh Kindarle - Free Press Release Website | Business Research