Facebook: Social Network King, Content Network Aspirant?

Some believe that Facebook’s recent privacy changes are an attempt to make more money by sharing this data with other businesses, or by presenting more targeted ads. But Facebook could have achieved the same by leaving privacy as it was, and just changing the Terms of Use to indicate that they could share the data with other businesses (for purposes of advertising).

I think Facebook’s motivation in moving to less private defaults is more ambitious. They don’t want to be just a keep-in-touch network, but also a content discovery network.

Content Discovery Networks Keep-in-touch Networks
Flickr, Twitter, Yelp Facebook, Email, IM
Optimized for propagation of content.
Goal is to allow best content
to surface to the top.
Optimized for non-public communication.
Users want their communication
to be widely read.
Users expect their communication
to only be readable by a trusted subset of users.
Allows formation of new connections with people you may not know in the real world,
allows one-way connections
Mirrors your offline relationships
Relationships, Content is public by default to make discovery easy Relationship and Content are not visible to everybody.

All social networks fall on one of the ends of a spectrum.

On one end are content discovery networks (Twitter, Flickr, Yelp). In these networks, engaged user communities share news, ideas and opinions. Content is public by default so it is easily discoverable. You can connect with people you don’t know in the real world. The focus is on enabling users to connect with content most relevant to them. Think of these as specialized search engines.

On the other side of the spectrum are (for lack of better word) keep-in-touch networks (Facebook, Email, IM). The purpose is communication between people as an end in itself. Connections mirror your real world relationships. There is an expectation that communication will not be public (only visible to people you trust). By this definition, Facebook, while being more public than Email/IM, is still a keep-in-touch network.

Keep-in-touch networks will always have more users than content discovery networks. There are far more people interested in non-public communication, than people interested in sharing  news/ideas/opinions. Even people who want to share ideas or opinions have a need for private communication. Which is why Twitter will never have as many users as Facebook. And yet, because content discovery networks play in the search space (they’re specialized search engines), I suspect content discovery networks are more profitable than keep-in-touch networks.

It’s hard to straddle both ends of the spectrum. Just ask Google, who tried adding a content discovery network (Buzz) to Gmail (which is a keep-in-touch network). Content discovery networks require more public defaults. Keep-in-touch networks require otherwise. It’s tricky to provide mechanisms for both types of communication.

And yet, Facebook is trying to do exactly that. They’ve always had some viral features typical of a content discovery network  (you can re-publish a status update from your friend as your own, you can become a “fan” of entities). But most users don’t use Facebook for content discovery – they use it to keep in touch with their social network.

Still, Facebook has kept at it. Facebook wants to be a content discovery network for web pages (like Twitter is). Make no mistake, the recent “Like” platform is an attempt to let users connect with entities that are relevant to them. Of course, to be successful, it needs to be viral, so one user can discover relevant entities (by looking at another person’s “likes”). This motivates the shift towards making a users “likes” public (which is also something Facebook did recently).

Will Facebook succeed? What do you think?

Update 11/1/2010: Dave McClure is saying something similar in his post titled: “How to Take down Facebook – Hint: It ain’t Twitter“.

Vertical Search will replace Search Ads

Imagine a search engine so good that it figures out your intent perfectly.

If you searched for “a upscale restaurant within 5 miles that’s open right now and serves steaks”, this mythical search engine’s top result would be exactly the one you want to go to. If you searched for “a radio controlled car available in a store within 10 miles”, it’s top result would be exactly the one you want to buy.

Such a search engine would never make money from search ads. The organic results would be so good, nobody would need to click on the ad. This is a critical insight: Search ads get clicks because the organic results are not perfect for some queries. If you believe that search engines will get better, you must concede that, all else being same, RoI from search ads will decrease.

The mythical search engine I talked about earlier – almost exists today. For my first query (about a restaurant), Yelp returns pretty good results. For my second query (a RC car), Milo returns pretty good results.  Truth is, for certain segments, companies have figured out how to present near-perfect search results to users. Kayak and Expedia do a good job for travel, Amazon and Google Product Search do it for products, and Yelp does it for local businesses. It’s also a growing market – Amazon’s latest quarterly profits were up 68% and Yelp’s traffic has increased 60% this last year. This growth comes at the expense of search ads.

Is this growth here to stay? I believe it is. And there are two reasons for it.

The first is just a generational shift: a new generation of web users is more web-savvy than the older one; the complex vertical search interfaces are less of a problem. The second is that the next surge of web usage will come from India and China – two countries where users are notoriously price-conscious. These “price optimizers” will not settle for an imprecise mechanism like search ads – they will want to express intent precisely, would want to compare and contrast offers from multiple businesses before making a choice. They will adopt vertical search interfaces at a faster rate, complex interfaces not withstanding.

This is not to say that search is going to die. Search is one of the most important technological problems of our age. But we’re going to have to find new ways to monetize it.

We’re all Islands of Technology Now

I was sitting at a restaurant today, alone, eating my dinner and reading news on my smartphone at the same time. And I realized that perhaps the smartphone has made me too efficient. Sure, it helps me search for good restaurants, gives me driving directions and lets me listen to music – but there are times (in the park, in the mall) where I previously would have sought out human interaction, but now just browse stuff on the smartphone.

It reminds me of something my father-in-law was telling me the other day. He remembers the days when telephones first became common – you’d have one land-line in a house, and if you talked on the phone with some relative, everybody congregated around the phone and you got to talk to everybody in the house. In its own way, the land-line was a more social experience. Then came cordless phones, and multiple handsets. Then came cell-phones. And now he rarely gets to talk to all the family members when he calls some relative – everybody has their own cell phone.

And smartphones just take it one step further. We’re even more cocooned in our own worlds now, at the expense of human interaction. I don’t think it’s completely a good thing. And yet, I’m not sure of the solution. Do we need classes to help us use technology wisely? Or maybe we just need data-plans that charge us by the byte. 🙂

Why Do People Click on Search Ads?

Google loses more than a dollar per month due to me.

A search query costs Google about 1 cent. I average about four queries a day, and almost never click on an ad (I’ve clicked on one search ad once in the last 8 years). I’m a net loss to Google. My search engine usage is subsidized by the vast sea of people who actually click on an ad.

In fact, I’m surprised that every user doesn’t behave like me.

I find search ads an inefficient mechanism to connect with service/product providers. Many times, the ads are not even relevant to my query – when I search for “android adoption europe”, the ads are about adopting children. Even when the ads are relevant, I am not sure if the business is legit, how the business was rated by other consumers and how does their product compare with other businesses’ products. In other words, search ads lack context. There’s always a better alternative to a search ad.

So then, why do people click on search ads?

I got my first clue when my brother-in-law (a medical doctor in the U.K.) visited me. He wanted to check his email, but instead of typing the URL, he Googled “doctors uk“, and then clicked on the first link. And he’s a well-educated, well-off person. My brother-in-law is not an exception – The top query on Google in 2006 was… “Yahoo”! Users use search engines as a mechanism to recall URLs.

Users find it too cumbersome to navigate to multiple web-sites (like Amazon or Yelp). It’s hard to remember and type in URLs (even harder on mobile devices). Indeed, I find myself clicking on ads far more often on mobile devices.  Others may find Amazon or Yelp’s search interface too complex. Many people are not even aware of specialized portals (did you know that you can find handicrafts at Etsy.com?). A combination of convenience and ignorance makes clicking on an ad an attractive proposition for many users.

But it’s getting easier to find the most relevant product or service online. Mobile devices have apps so you don’t have to type URLs. Vertical search engine like Ebay, Yelp and Amazon are becoming more popular. And what’s better, as they become popular, they rank higher in the top 10 results on search engines – so users don’t have to remember URLs of other sites and don’t have to click on ads either. All these trends put a downward pressure on search RoI.

I believe that these specialized, vertical portals will eventually replace the search ads model. But that is a topic for another post.

Are Hyperlinks losing relevance as a measure of relevance?

The web started off as a network of hyper-links, which is what made PageRank (and Google) so effective. The assumption that important content linked to other important content, enabled Google to sift relevant content right to the top. And it worked well for a long time.

But a subtle shift is underway. Increasingly, relevance of content on the web is being evaluated using other techniques. Amazon (and Yelp) rank products (local businesses, in case of Yelp) by having users review them. Even the reviews are ranked by user voting. There are other sites – Stack Overflow, Seeking Alpha and others, which also use a combination of user voting and editorial control.

Think about it: Why aren’t these sites using hyperlinks to connect relevant content with other relevant content? Surely, doing so has its advantages (search engine algorithms could automatically rank content on the site).

There are three reasons why the long term trend is against hyperlinks (as a relevance indicator):

  1. Content is being added to the web at a high enough rate that publishers don’t have the time and resources to seek out other relevant content and link to it. Hyper-links are too heavyweight a mechanism.
  2. A hyperlink-based system requires content publishers to expend extra effort to find other relevant content and link to it; it’s more efficient to have the readers of content expend extra effort. After all, there are many more readers of content than content publishers. Hence the move towards user voting, likes, retweets, etc.
  3. And lastly, because these techniques use the human in the loop more effectively, they are more efficient (in terms of cloud costs) than purely algorithmic approaches (like PageRank).

We’ll see more of this in the coming years. Content will be ranked using engaged user communities. Hyperlinks will be less important in deciding relevance.

Post Script: If this all seems obvious to you, then consider this: many pundits consider Activity Streams to be the logical successor to hyperlinks, yet it suffers from some of the same problems. Click here for details.

Information wants to be Organized

Fifteen years ago, Yahoo attempted to classify all of the web into a heirarchical directory. That attempt failed.

The web grew to be too big and too decentralized for a heirarchical approach to work. In the world of blogs and small websites, and the long tail of content, Google’s algorithmic approach to surfacing relevant information won out.

And yet, it’s been interesting to watch an increasing trend towards centralization in the last few years. There’s Wikipedia for facts. Yelp.com for local businesses (and local business reviews). There’s SeekingAlpha.com for opinions on financial matters. There’s Amazon.com for products (and product reviews).

I call this organized information.

So what happened?

What happened is that there is just too much damn content out there now. In a world of too much content, an effective search experience is necessary. When semantically related information is organized into a portal, a deep search experience can be provided to users that leverages content semantics. For example, Yelp’s search experience is tailored to its content (you can filter by location, by type of business, by quality of the reviewer, etc).

The second advantage, for content publishers (on say, SeekingAlpha) is that it increases their chances of being heard. With so much content on the web today, it is no longer sufficient to just have something interesting to say. A dedicated portal for a content category attracts more eyeballs, attracts more hyperlinks from other authorities and a higher rank on search engines. A dedicated portal also has an engaged community that can filter relevant content to the top faster.

This trend will continue.

Who knows, in the next 15 years, we may yet end up with a self-organized directory of content on the web. And we may realize that Yahoo was really an idea ahead of its time.

Related Post: “Are Hyperlinks losing relevance as a measure of relevance?

Google, Watch out to the “Mac Minority”

I find the search business model strange.

What I find strange is that I am a fairly heavy search engine consumer, reasonably well-off, and yet I pay no money to the search engine. I have never clicked on a search ad. My search usage is subsidized by the vast sea of people who actually click on an ad.

And I am not alone. I am sure if I went to Google (the corporation) and asked them how they connected with products and services – I’d hear that they used Amazon.com, Ebay.com or Yelp.com or some other site. Anything but ads. It’s as if the engineering team of Windows had decided to use Macs.

My conjecture is that the people who don’t click on ads are a part of a small minority – I call them the “mac minority” (I’ll explain in a sec) – who use the search engine heavily, and yet are savvy enough to never click on an ad. This minority knows how to dig up information more effectively than a Google ad. This “mac minority” is also well-educated, more well-off than general population and heavier users of search – in other words, an attractive demographic for search engine companies. And yet, Google makes little money out of this demographic.

Hopefully, Google is paying attention to this fact. Most definitely, investors should pay attention to this fact. Because behind this fact lies a cautionary tale.

The “Mac Minority”, explained

Flashback to the year 2000. Microsoft’s Windows had a huge market share of the computing devices market. Apple’s Mac computers had a tiny market share – less than 2%. I call these Mac users a “mac minority”. This “mac minority” was well-off, educated and discriminating, and Apple consciously went after these users. For a while, Windows kept winning (because of its cheaper price, application ecosystem, its focus on enterprise, etc). But within 10 years, we have reached an inflection point (the mobile device revolution) that allowed more and more consumers to afford an Apple device – in other words, to behave like the aforementioned minority. And just like that, Apple now has more growth than Microsoft.

Ignore the “Mac Minority” at your peril

If the “mac minority” of the search user-base prefers to use something other than Google ads, we need to ask why that is the case (and why the rest of the users don’t). Maybe the majority of users don’t know that Amazon.com exists. Maybe they do know of Amazon.com‘s existence but find it cumbersome to type the URL. Maybe users find Amazon.com‘s user-interface confusing. I don’t know.

What I do know is that eventually these barriers will go away. In 10 years, a new generation of web-savvy internet users will be aware of Amazon.com. They will find it easy to type its URL. They will find the Amazon.com user-interface intuitive.

Bottomline: if the “mac minority” of the search segment is a hold-out, then Google should watch out for an inflection point that makes the majority of users behave like that minority. Because it will happen. Google has to focus on monetizing the “mac minority”, because otherwise the search ads business may be gone faster than they think.